Initial public upload from Projekte/AgentProjectMinenraeumer (part 9)

.gitattributes

@@ -88,3 +88,5 @@ ProjectIntakeTower/explainers/video/sympy__sympy-20590-2a70debb9a.mp4 filter=lfs
 ProjectIntakeTower/explainers/video/sympy__sympy_c5cef2499d6eed024b0db5c792d6ec7c53baa470-10b468569f.mp4 filter=lfs diff=lfs merge=lfs -text
 ProjectIntakeTower/explainers/video/test-examples-62e65bab8c.mp4 filter=lfs diff=lfs merge=lfs -text
 ProjectIntakeTower/explainers/video/v1_open_claw_build_liteversion_agent_star-7728171bc8.mp4 filter=lfs diff=lfs merge=lfs -text
+ProjectIntakeTower/understanding_site/site-data.js filter=lfs diff=lfs merge=lfs -text
+ProjectIntakeTower/visual_summary/project_learning_graphic.png filter=lfs diff=lfs merge=lfs -text

ProjectIntakeTower/inventory/project_cards/samuel-system-1690f00467.md

@@ -0,0 +1,29 @@
+# samuel-system
+
+- Project ID: `samuel-system-1690f00467`
+- Path: `/home/smlflg/Projekte/LuckSurface/samuel-system`
+- Last modified: `2026-04-21T15:52:06+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `6` — README present
+- Evidence: `inventory/evidence/samuel-system-1690f00467.json`
+
+## README summary
+
+# Samuel.SYSTEM Samuel.SYSTEM is the public operating model for Samuel's agentic work. It connects: - private focus and work state from `samuel-os-v2` - runtime experiments from OpenClaw - proof artifacts from PortfolioTimeline - publishing drafts from Paperclip - public positioning through `smlfg.github.io`
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sandraswritingworld-9c951f9f13.md

@@ -0,0 +1,38 @@
+# SandrasWritingWorld
+
+- Project ID: `sandraswritingworld-9c951f9f13`
+- Path: `/home/smlflg/Projekte/SandrasWritingWorld`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: docs/notes
+- Triage status: `active`
+- Risk level: `low`
+- Score: `40` — git repo, 31 TODO/FIXME/WIP hints, README present, plans/backlogs found
+- Evidence: `inventory/evidence/sandraswritingworld-9c951f9f13.json`
+
+## README summary
+
+# Your Website — SandrasWritingWorld Welcome! This is your personal website. Everything you write here — poems, stories, essays, events — lives at **sandraswritingworld.netlify.app**. This guide tells you everything you need to know to manage it yourself. No coding required. --- ## How to Add a New Text 1. Go to **sandraswritingworld.netlify.app/admin** 2. Click **Login** and use your email address 3. In the left menu, choose a category:
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:101 — - Pixel-art board colors via cm-chessboard CSS overrides (black + crimson or black + gold — open question)
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:163 — **No build step.** Pure static. Open `index.html` in a browser to develop locally.
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:350 — ## 9. Open Questions for Sandra / Miran
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:371 — - [ ] Navigate between all 6 pages: SoundCloud player never stops or restarts
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:372 — - [ ] Home: U-Bahn video autoplays muted, loops, fills viewport
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:373 — - [ ] Home: "Welcome to the journey" overlay text visible over video
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:374 — - [ ] Chess: playable game against AI at all 3 difficulties
+- `/home/smlflg/Projekte/SandrasWritingWorld/ARCHITECTURE.md`:375 — - [ ] Chess: chess scripts only load when navigating to `#/chess` (check Network tab)
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/schema-e571b59e0e.md

@@ -0,0 +1,30 @@
+# schema
+
+- Project ID: `schema-e571b59e0e`
+- Path: `/home/smlflg/Projekte/Harness_swap/packages/schema`
+- Last modified: `2026-04-23T05:44:11+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: node
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/schema-e571b59e0e.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/schloss_pdf-9fe7072a7c.md

@@ -0,0 +1,33 @@
+# schloss_PDF
+
+- Project ID: `schloss_pdf-9fe7072a7c`
+- Path: `/home/smlflg/Projekte/schloss_PDF`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `unknown`
+- Technologies: unknown
+- Triage status: `review`
+- Risk level: `low`
+- Score: `5` — git repo, 2 TODO/FIXME/WIP hints, plans/backlogs found, unclear purpose
+- Evidence: `inventory/evidence/schloss_pdf-9fe7072a7c.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/schloss_PDF/OpenCode PDF Zusammenfassung_files/writing-block-editor-o0v3uy9j.css`:1 — ._4ccM4G_magic-edit-selection{background:var(--theme-user-selection-bg,var(--selection));-webkit-box-decoration-break:clone;box-decoration-break:clone;padding:4.5px 0}[data-me-role=container]{display:block;position:relative}.writing-block-editor.magic-edit-min
+- `/home/smlflg/Projekte/schloss_PDF/.hermes/plans/2026-04-07_213045-bednke-was-dass-emotianl-f-r-mich-bedeutet.md`:91 — **Opus 4.6:** Principle-first, Review-first, Senior-Analyst-Frame
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scim-8ee0630682.md

@@ -0,0 +1,30 @@
+# scim
+
+- Project ID: `scim-8ee0630682`
+- Path: `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/proxy/management_endpoints/scim`
+- Last modified: `2026-03-23T15:16:26+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `8` — 1 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/scim-8ee0630682.json`
+
+## README summary
+
+# SCIM v2 Integration for LiteLLM Proxy This module provides SCIM v2 (System for Cross-domain Identity Management) endpoints for LiteLLM Proxy, allowing identity providers to manage users and teams (groups) within the LiteLLM ecosystem. ## Overview SCIM is an open standard designed to simplify user management across different systems. This implementation allows compatible identity providers (like Okta, Azure AD, OneLogin, etc.) to automatically provision and deprovision users and groups in LiteLLM Proxy. ## Endpoints The SCIM v2 API follows the standard specification with the following base URL: ``` /scim/v2
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/proxy/management_endpoints/scim/README_SCIM.md`:7 — SCIM is an open standard designed to simplify user management across different systems. This implementation allows compatible identity providers (like Okta, Azure AD, OneLogin, etc.) to automatically provision and deprovision users and groups in LiteLLM Proxy.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/screenshots-4876cc00af.md

@@ -0,0 +1,30 @@
+# screenshots
+
+- Project ID: `screenshots-4876cc00af`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/docs/public/screenshots`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/screenshots-4876cc00af.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-12ad857658.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-12ad857658`
+- Path: `/home/smlflg/.hermes/profiles/dlm/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-12ad857658.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-18a710a12f.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-18a710a12f`
+- Path: `/home/smlflg/.hermes/profiles/smk/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-18a710a12f.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-57575003f0.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-57575003f0`
+- Path: `/home/smlflg/.hermes/profiles/fixer/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-57575003f0.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-5e89206969.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-5e89206969`
+- Path: `/home/smlflg/.hermes/profiles/voice-learning/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-5e89206969.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-617f385573.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-617f385573`
+- Path: `/home/smlflg/.hermes/profiles/decision-control/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-617f385573.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-b70185e298.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-b70185e298`
+- Path: `/home/smlflg/.hermes/profiles/ra/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-b70185e298.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-e206d00442.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-e206d00442`
+- Path: `/home/smlflg/.hermes/profiles/diagnosis/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-e206d00442.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-f22bd304ed.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-f22bd304ed`
+- Path: `/home/smlflg/.hermes/profiles/openclaude-mcp/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-f22bd304ed.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/scripts-f979721f5a.md

@@ -0,0 +1,30 @@
+# scripts
+
+- Project ID: `scripts-f979721f5a`
+- Path: `/home/smlflg/.hermes/profiles/mlg/skills/productivity/google-workspace/scripts`
+- Last modified: `2026-04-27T08:42:52+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — 
+- Evidence: `inventory/evidence/scripts-f979721f5a.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sdists-v9-be92bdb037.md

@@ -0,0 +1,30 @@
+# sdists-v9
+
+- Project ID: `sdists-v9-be92bdb037`
+- Path: `/home/smlflg/.hermes/.uv-cache/sdists-v9`
+- Last modified: `2026-03-23T15:16:33+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `unknown`
+- Technologies: unknown
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — git repo, unclear purpose
+- Evidence: `inventory/evidence/sdists-v9-be92bdb037.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sdists-v9-c396947c6e.md

@@ -0,0 +1,30 @@
+# sdists-v9
+
+- Project ID: `sdists-v9-c396947c6e`
+- Path: `/home/smlflg/Projekte/Connector/google-workspace/state/cache/uv/sdists-v9`
+- Last modified: `2026-04-28T06:11:15+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `unknown`
+- Technologies: unknown
+- Triage status: `archive_candidate`
+- Risk level: `low`
+- Score: `0` — git repo, unclear purpose
+- Evidence: `inventory/evidence/sdists-v9-c396947c6e.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/searchagent-7ebc193f7e.md

@@ -0,0 +1,30 @@
+# SearchAgent
+
+- Project ID: `searchagent-7ebc193f7e`
+- Path: `/home/smlflg/Projekte/SearchAgent`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `review`
+- Risk level: `low`
+- Score: `16` — git repo, 1 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/searchagent-7ebc193f7e.json`
+
+## README summary
+
+# SearchAgent Archived learning project. Built to understand the "Search" layer of Search + RAG pipelines. ## What it taught - How search engines return structured results (title, URL, snippet) - How to pass search results as context into an LLM prompt - Difference between raw HTML and structured search results ## Components - `search.py` — DuckDuckGo search via `ddgs`, returns `{title, url, snippet}`
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/SearchAgent/FOR_SMLFLG.md`:53 — ## Next Steps (wenn weiterbauen)
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/self_investment-9cc895ec7e.md

@@ -0,0 +1,38 @@
+# self_investment
+
+- Project ID: `self_investment-9cc895ec7e`
+- Path: `/home/smlflg/Projekte/Amazon/self_investment`
+- Last modified: `2026-05-03T18:43:57+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `32` — 13 TODO/FIXME/WIP hints, README present, plans/backlogs found
+- Evidence: `inventory/evidence/self_investment-9cc895ec7e.json`
+
+## README summary
+
+# Self-Investment Operating Plan Start date: 2026-05-04 This folder turns the 15k net self-investment plan into an executable routine. It is intentionally simple: one budget CSV, one task CSV, one weekly review, and one tracker script. ## Source of truth Money truth: 1. Broker transactions and bank balances. 2. Tax reserve calculated from realised gains. Use your real tax situation, not this model, for the final number. 3. `budget.csv` for the plan allocation.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Amazon/self_investment/tracker.py`:23 — with path.open(newline="", encoding="utf-8") as handle:
+- `/home/smlflg/Projekte/Amazon/self_investment/tracker.py`:28 — with path.open("w", newline="", encoding="utf-8") as handle:
+- `/home/smlflg/Projekte/Amazon/self_investment/tracker.py`:98 — print(f"Next {args.limit} tasks")
+- `/home/smlflg/Projekte/Amazon/self_investment/tracker.py`:135 — next_tasks = sub.add_parser("next", help="Show next open tasks")
+- `/home/smlflg/Projekte/Amazon/self_investment/weekly_review.md`:1 — # Weekly Review
+- `/home/smlflg/Projekte/Amazon/self_investment/weekly_review.md`:33 — Next week's 3 tasks:
+- `/home/smlflg/Projekte/Amazon/self_investment/weekly_review.md`:69 — Next week's 3 tasks:
+- `/home/smlflg/Projekte/Amazon/self_investment/README.md`:5 — This folder turns the 15k net self-investment plan into an executable routine. It is intentionally simple: one budget CSV, one task CSV, one weekly review, and one tracker script.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/server-8bd5de99fa.md

@@ -0,0 +1,38 @@
+# server
+
+- Project ID: `server-8bd5de99fa`
+- Path: `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server`
+- Last modified: `2026-04-21T13:03:27+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: node, typescript
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `24` — 12 TODO/FIXME/WIP hints
+- Evidence: `inventory/evidence/server-8bd5de99fa.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/logicParser.ts`:95 — const next = nonEmpty[i + 1];
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/logicParser.ts`:96 — const nextIndent = next?.raw.match(/^ */)?.[0].length ?? -1;
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/logicParser.ts`:97 — if (!next || nextIndent <= indent) {
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/fsrs.ts`:50 — review: log.review.toISOString(),
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/fsrs.ts`:75 — const result = scheduler.next(previous, new Date(), ratingFromGrade(grade));
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/fsrs.ts`:79 — next: serializeCard(result.card),
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/routes/index.ts`:591 — db.prepare("UPDATE attempts SET status = 'completed', current_stage = 'review', completed_at = ?, updated_at = ? WHERE id = ?")
+- `/home/smlflg/Projekte/PseudoLogikPyhthonCPP/server/src/routes/index.ts`:597 — api.get('/review/due', (c) => {
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Classify purpose from filenames only; ask owner before writing project README.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/session-browser-257de3e006.md

@@ -0,0 +1,35 @@
+# session-browser
+
+- Project ID: `session-browser-257de3e006`
+- Path: `/home/smlflg/Projekte/MyAIGame/session-browser`
+- Last modified: `2026-02-28T12:07:24+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: python
+- Triage status: `review`
+- Risk level: `low`
+- Score: `10` — 5 TODO/FIXME/WIP hints
+- Evidence: `inventory/evidence/session-browser-257de3e006.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/MyAIGame/session-browser/session_parser.py`:153 — with open(filepath, "r", encoding="utf-8", errors="replace") as f:
+- `/home/smlflg/Projekte/MyAIGame/session-browser/session_parser.py`:163 — with open(filepath, "r", encoding="utf-8", errors="replace") as f:
+- `/home/smlflg/Projekte/MyAIGame/session-browser/session_parser.py`:272 — with open(path, "r", encoding="utf-8", errors="replace") as f:
+- `/home/smlflg/Projekte/MyAIGame/session-browser/session_parser.py`:442 — with open(path, "r", encoding="utf-8", errors="replace") as f:
+- `/home/smlflg/Projekte/MyAIGame/session-browser/session_parser.py`:458 — # Gather context lines (previous + next)
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Classify purpose from filenames only; ask owner before writing project README.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/shared-66bb41fa0c.md

@@ -0,0 +1,33 @@
+# shared
+
+- Project ID: `shared-66bb41fa0c`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/shared`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: node, typescript
+- Triage status: `review`
+- Risk level: `low`
+- Score: `5` — 2 TODO/FIXME/WIP hints, plans/backlogs found
+- Evidence: `inventory/evidence/shared-66bb41fa0c.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/shared/data/sample-documents/meeting-summary.txt`:16 — 4. Next Steps
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/shared/src/renderer/App.tsx`:52 — // In a real app this would open a file dialog.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/shared-gui-39cb40d2b3.md

@@ -0,0 +1,32 @@
+# shared-gui
+
+- Project ID: `shared-gui-39cb40d2b3`
+- Path: `/home/smlflg/Projekte/shared-gui`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `19` — git repo, 2 TODO/FIXME/WIP hints, README present, plans/backlogs found
+- Evidence: `inventory/evidence/shared-gui-39cb40d2b3.json`
+
+## README summary
+
+# shared-gui — GTK3 Base Class Shared GTK3 base for all desktop apps in this setup. Provides `BaseApp`: a themed GTK3 window with Catppuccin + COSMIC dark/light auto-detection. ## What BaseApp provides - GTK3 window with HeaderBar, content area, StatusBar - Catppuccin theme (auto-detects COSMIC dark/light mode, live-updates on theme switch) - Helper methods: `set_subtitle()`, `add_header_widget()`, `set_status()`, `start_refresh()`, `add_css()` - CSS classes: `.base-card`, `.base-toolbar`, `.base-badge`, `.base-badge-ok/warn/error`, `.base-accent-btn`
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/shared-gui/FOR_SMLFLG.md`:70 — - Kein Packaging/`__init__.py` — direkter Import via sys.path Hack
+- `/home/smlflg/Projekte/shared-gui/FOR_SMLFLG.md`:81 — ## Next Steps
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/shortcutcounter-d2ffa4d801.md

@@ -0,0 +1,39 @@
+# ShortcutCounter
+
+- Project ID: `shortcutcounter-d2ffa4d801`
+- Path: `/home/smlflg/Projekte/ShortcutCounter`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: dirty (2 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `51` — git repo, tracked dirty work, 8 TODO/FIXME/WIP hints, README present, plans/backlogs found
+- Evidence: `inventory/evidence/shortcutcounter-d2ffa4d801.json`
+
+## README summary
+
+# ShortcutCounter Passive keyboard shortcut tracker for Linux/Wayland. Runs as a background daemon, records only modifier+key combinations (e.g. `Ctrl+C`), and shows usage stats via CLI or the ClaudeCodePanel tab. ## Installation ```bash # 1. Install dependencies pip install evdev click rich # 2. Add yourself to the input group (required for evdev access) sudo usermod -aG input $USER
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:5 — Monster-Session: 5-Agent Team Sprint (openclaw-sprint) mit Research, Architektur, Builder, Tester, Coach. Danach /simplify Review mit 3 Fixes. Session-Feedback: Sidecar Suggestions brauchen Phase-Awareness.
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:10 — - [ ] ShortcutCounter committen: `storage.py` (dead tomli fallback entfernt), `services/shortcut-counter.service` (StartLimit → [Unit]), `pyproject.toml` (toml dep entfernt)
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:11 — - [ ] OpenClaw docs committen: 10 neue READMEs in `~/opencode/openclaw/docs-samuel/` + 3 Fixes (Kommentar, Cron-Snippet, PRIOR_ART.md)
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:14 — - [ ] Telefonnummer in `~/opencode/openclaw/HOWTO-SAMUEL.md` Zeile 257 — rauslöschen oder lassen?
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:17 — - [ ] Sidecar Suggestion-Logik phase-aware machen:
+- `/home/smlflg/Projekte/ShortcutCounter/PLAN_2026-03-05_sprint-cleanup.md`:24 — - [ ] /learn — Agent Teams Workflow verinnerlichen (was hat Samuel verstanden?)
+- `/home/smlflg/Projekte/ShortcutCounter/FOR_SMLFLG.md`:107 — - `get_today()` Bug: Alte Version zeigte ALL-TIME count gefiltert nach last_used. Fix: eigene daily_count/daily_date Spalten mit auto-reset
+- `/home/smlflg/Projekte/ShortcutCounter/shortcut_counter/storage.py`:53 — with open(p, "rb") as f:
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Read plan/backlog artifacts and extract current unfinished thread.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidebyside-050b659506.md

@@ -0,0 +1,37 @@
+# sidebyside
+
+- Project ID: `sidebyside-050b659506`
+- Path: `/home/smlflg/Projekte/HarnessEcosystem/sidebyside`
+- Last modified: `2026-04-23T05:40:34+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `22` — 8 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/sidebyside-050b659506.json`
+
+## README summary
+
+# Sidebyside Harness Snapshot Standalone statische SPA fuer `data/harness-snapshot.json`. ## Starten Die Seite laedt die JSON-Datei per `fetch`, deshalb braucht sie einen lokalen HTTP-Server: ```bash python3 -m http.server 8080 ``` Dann im Browser: `http://localhost:8080`
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/app.js`:567 — ui.dialog.showModal ? ui.dialog.showModal() : ui.dialog.setAttribute("open", "");
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/app.js`:572 — if (ui.dialog.open) ui.dialog.close();
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_hermes.py`:391 — warnings.append(f"unable to open state db {path}: {exc}")
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_claude.py`:214 — with path.open("r", encoding="utf-8", errors="replace") as handle:
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_claude.py`:255 — with file_path.open("r", encoding="utf-8", errors="replace") as handle:
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_codex.py`:201 — with path.open("rb") as fh:
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_codex.py`:232 — warnings.append(f"Failed to open sqlite database {path}: {exc}")
+- `/home/smlflg/Projekte/HarnessEcosystem/sidebyside/scripts/extract_codex.py`:340 — with path.open("r", encoding="utf-8") as fh:
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-93e2e4592e.md

@@ -0,0 +1,39 @@
+# Sidecar
+
+- Project ID: `sidecar-93e2e4592e`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar`
+- Last modified: `2026-05-04T17:33:23+00:00`
+- Git repo: yes
+- Git status: dirty (1 tracked changes)
+- Branch: `master`
+- Technologies: docs/notes
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `63` — git repo, tracked dirty work, 40 TODO/FIXME/WIP hints, README present, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-93e2e4592e.json`
+
+## README summary
+
+# Sidecar v5 — Quality Control for Claude Code A persistent daemon that monitors Claude Code in real-time and injects situational context. Five layers of quality control: 1. **Pattern Detection** (deterministic, $0, ~1ms) — Detects anti-patterns, loops, stalls, error cascades 2. **Quick Judge** (LLM, ~$0.0002/call, SYNC) — Evaluates work quality against coaching rules 3. **Deep Judge Agent** (LLM + function calling, ~$0.005/call, ASYNC) — Autonomous coach that navigates a knowledge base and gives personalized feedback 4. **Learning & Trends** — Tracks implemented vs learned gaps, cross-session patterns, long-term trends 5. **Eval Pipeline** (offline) — EventStore, 30 scenarios, A/B testing, drift detection, rule economics ## Architecture
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/content_detectors.py`:34 — with open("/tmp/claude-sidecar.log", "a") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/florian_tts.py`:35 — with open(LOG_PATH, "a") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/florian_tts.py`:179 — with open(AUDIO_LOCK_PATH, "a+", encoding="utf-8") as lock_file:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/context_selector.py`:134 — with open("/tmp/claude-sidecar.log", "a") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/sidecar.py`:140 — with open(LOG_PATH, "a") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/sidecar.py`:196 — "keywords": ["fix", "nur hier", "subsystem", "bug"],
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/sidecar.py`:417 — with open(OVERRIDES_PATH, "r") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar/sidecar.py`:574 — with open(OVERRIDES_PATH, "r") as f:
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Read plan/backlog artifacts and extract current unfinished thread.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-ng-6382aa0e95.md

@@ -0,0 +1,38 @@
+# Sidecar-ng
+
+- Project ID: `sidecar-ng-6382aa0e95`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (6 tracked changes)
+- Branch: `master`
+- Technologies: docs/notes
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `59` — git repo, tracked dirty work, 40 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/sidecar-ng-6382aa0e95.json`
+
+## README summary
+
+# Sidecar-NG Stateless Context Injection Hook for Claude Code. Ein Script, kein Daemon. ## Was ist Sidecar-NG? Sidecar-NG analysiert jeden User-Prompt via Claude Code Hook (`UserPromptSubmit`) und injiziert **eine** situativ relevante Regel als `additionalContext`. Ergebnis von 8 Iterationen (V1-V8, 45K+ LOC Peak) — reduziert auf ~1.200 LOC mit besserer Accuracy. ## Architektur ``` UserPromptSubmit (stdin: JSON) │
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_results.json`:225 — "category": "SKILL: review",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_results.json`:226 — "expected": "SKILL: review",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_results.json`:227 — "got_rule": "WICHTIG: Lies zuerst ~/Projekte/Sidecar-ng/knowledge/commands/review.md und befolge den Workflow.",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_results.json`:230 — "prompt_short": "mach ein code review von den heutigen änderungen"
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_ab_results.json`:227 — "category": "SKILL: review",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_ab_results.json`:228 — "expected": "SKILL: review",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_ab_results.json`:229 — "got_rule": "WICHTIG: Lies zuerst ~/Projekte/Sidecar-ng/knowledge/commands/review.md und befolge den Workflow.",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng/eval_ab_results.json`:232 — "prompt_short": "mach ein code review von den heutigen änderungen"
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-ng-dashboard-8736c7518e.md

@@ -0,0 +1,39 @@
+# Sidecar-ng-dashboard
+
+- Project ID: `sidecar-ng-dashboard-8736c7518e`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `unknown`
+- Technologies: agent, python
+- Triage status: `active`
+- Risk level: `low`
+- Score: `50` — git repo, 15 TODO/FIXME/WIP hints, README present, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-ng-dashboard-8736c7518e.json`
+
+## README summary
+
+# Sidecar-NG Dashboard A standalone web dashboard for monitoring and managing Sidecar-NG, a prompt-classifier with keyword-first pre-filtering and LLM-fallback chaining. ## Overview The Sidecar-NG Dashboard provides real-time visibility into: - **Prompt injections** — Track all classifier decisions and LLM calls - **Safety gates** — Monitor gate decisions (allow/deny) for tool executions - **Knowledge base** — View and edit priority knowledge files - **Provider control** — Enable or disable LLM providers from the dashboard
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:91 — "defaults": ["Repository scope", "Bug and regression review"],
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:121 — "defaults": ["Current topic", "Open research question"],
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:250 — current = next(provider for provider in providers if provider["name"] == provider_name)
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:280 — updated = next(
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:336 — dashboard_session = next(
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/control_service.py`:385 — updated = next(mode for mode in list_modes() if mode["name"] == mode_name)
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/main.py`:275 — with open(INJECT_LOG) as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-ng-dashboard/main.py`:283 — with open(INJECT_LOG) as f:
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v5-ee6c377f82.md

@@ -0,0 +1,38 @@
+# Sidecar-V5
+
+- Project ID: `sidecar-v5-ee6c377f82`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (1 tracked changes)
+- Branch: `main`
+- Technologies: python
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `58` — git repo, tracked dirty work, 12 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/sidecar-v5-ee6c377f82.json`
+
+## README summary
+
+# Sidecar V5 — Personal Context Awareness Standalone asyncio daemon that injects personal context about Samuel into Claude Code. Knows Samuel's ADHD profile, work habits, and mood — gives neutral, timely recommendations. **Not coupled to Sidecar V1-V4.** Reads their session files but runs independently. ## What it does - Loads Samuel's personality profiles (ChatGPT + Gemini analysis, NotebookLM, Subconsciousness) - Detects patterns: late-night sessions, no-break hyperfocus, session sprawl, mood shifts - Injects max 2 recommendations per prompt via Claude Code's UserPromptSubmit hook
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/sidecar_v5.py`:137 — Fail-open: bei jedem Fehler {} zurueckgeben.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/sidecar_v5.py`:166 — # IRON RULE: Never crash. Fail-open.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/v5_bridge.py`:8 — Fail-open: bei jedem Fehler leeren Output, Claude arbeitet weiter.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/v5_bridge.py`:74 — # Daemon nicht erreichbar oder Fehler → fail-open
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/FOR_SMLFLG.md`:70 — - Asyncio Unix Socket Server mit Fail-Open (jeder Fehler → leere Antwort, kein Crash)
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/FOR_SMLFLG.md`:89 — - **Fail-Open als Iron Rule**: Jeder try/except gibt `{}` zurueck. Profil-Laden fehlgeschlagen = degraded mode, kein Absturz.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/FOR_SMLFLG.md`:95 — ## Next Steps
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V5/README.md`:13 — - Never blocks. Fail-open (returns empty if something goes wrong)
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v6-bf4e8dc3cb.md

@@ -0,0 +1,39 @@
+# Sidecar-V6
+
+- Project ID: `sidecar-v6-bf4e8dc3cb`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: agent, python
+- Triage status: `active`
+- Risk level: `low`
+- Score: `56` — git repo, 40 TODO/FIXME/WIP hints, README present, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-v6-bf4e8dc3cb.json`
+
+## README summary
+
+# Sidecar V6 AI Coaching Companion fuer Claude Code — beobachtet deine Coding-Sessions und gibt ADHS-aware Nudges, trackt Mechanismen mit Voting, und hat einen MiniMax Slow-Path fuer tiefere Analyse. Mit Florian (TTS) spricht er dir auch zu. ## Quick Start ```bash # 1. Virtualenv einrichten cd ~/Projekte/Sidecar-V6
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/PLAN_2026-03-09_AGENTISCHE-MECHANISMEN.md`:150 — ## Fail-Open
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/architektur-v6.html`:1342 — <div class="mech-desc">Blockiert neue Projekte wenn aktive unfertig sind. WIP-Limit.</div>
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/architektur-v6.html`:1359 — <div class="mech-name">Weekly Review</div>
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/architektur-v6.html`:1458 — <div class="mech-name">Review-Vorschlag</div>
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/TODO_dashboard.md`:1 — # TODO: Dashboard + Panel End-to-End verdrahten
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/TODO_dashboard.md`:75 — - [ ] `/api/xp` returned XP-Bar Partial mit Level, XP today, Streak
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/TODO_dashboard.md`:76 — - [ ] `/api/findings` returned Findings-Liste
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V6/TODO_dashboard.md`:77 — - [ ] `/api/mechanism/{id}` returned Detail-View
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v7-ed034c870d.md

@@ -0,0 +1,40 @@
+# Sidecar-V7
+
+- Project ID: `sidecar-v7-ed034c870d`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (5 tracked changes)
+- Branch: `main`
+- Technologies: agent, python
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `77` — git repo, tracked dirty work, 40 TODO/FIXME/WIP hints, README present, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-v7-ed034c870d.json`
+
+## README summary
+
+# Sidecar V7 **Context-aware ADHS coaching daemon for AI coding assistants.** ## Why Developers with ADHD face specific challenges when coding with AI assistants: hyperfocus without breaks, session sprawl across too many projects, and no awareness of energy patterns. Sidecar observes tool events in real-time and coaches through contextual nudges, never blocking the workflow. Everything is fail-open — if the daemon goes down, your coding assistant works exactly as before. ## Architecture ``` Claude Code Hooks (bridge.py) -> Unix Socket /tmp/sidecar-v7.sock
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/FOR_SMLFLG.md`:36 — Jede Quelle ist fail-open: wenn eine Quelle ausfaellt, funktioniert der Rest weiter.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/FOR_SMLFLG.md`:50 — | V6 core/ | daemon.py, coach.py, ... | core/sidecar_core/ | v6→v7 Rename, Bug-Fixes |
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/FOR_SMLFLG.md`:116 — - Fail-Open ist Pflicht: Daemon darf Claude Code nie blockieren
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/FOR_SMLFLG.md`:125 — - `_check_should_speak()` — Neue Methode die `tts_ranker.should_speak()` aus `samuel/sidecar_samuel/tts_ranker.py` verdrahtet. Fail-open Gate: wenn eine Mechanismus+msg_type+phase+hour_bucket Kombination in der DB schlecht bewertet wurde (Score < 0.45), wird S
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/FOR_SMLFLG.md`:149 — **Fail-open Gate:** Wenn keine Daten da sind oder DB-Error → erlauben. Nur blockieren wenn genug negative Bewertungen vorliegen. Sicherer Default weil ein stummes System nervoser macht als ein leicht zu geschwätziges.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/README.md`:11 — Developers with ADHD face specific challenges when coding with AI assistants: hyperfocus without breaks, session sprawl across too many projects, and no awareness of energy patterns. Sidecar observes tool events in real-time and coaches through contextual nudg
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/README.md`:56 — # Open dashboard
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7/README.md`:77 — | `adhs` | 0 (Health) | Hyperfocus guard, big-bang guard, zombie bug detector |
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Read plan/backlog artifacts and extract current unfinished thread.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v7-lite-323322aa5b.md

@@ -0,0 +1,36 @@
+# Sidecar-V7-Lite
+
+- Project ID: `sidecar-v7-lite-323322aa5b`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-Lite`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (2 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `needs_human`
+- Risk level: `medium`
+- Score: `37` — git repo, tracked dirty work, 4 TODO/FIXME/WIP hints, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-v7-lite-323322aa5b.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-Lite/PLAN_V7_LITE_M2.7.md`:266 — - and the agent type is long-running or open-ended
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-Lite/sidecar_lite/bridge.py`:68 — # Fail-open: any error means we don't block Claude
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-Lite/sidecar_lite/session.py`:119 — with open(path, "w") as f:
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-Lite/sidecar_lite/session.py`:128 — with open(path) as f:
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Read plan/backlog artifacts and extract current unfinished thread.
+4. Classify purpose from filenames only; ask owner before writing project README.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v7-v2-520073a262.md

@@ -0,0 +1,39 @@
+# Sidecar-V7-V2
+
+- Project ID: `sidecar-v7-v2-520073a262`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `active`
+- Risk level: `low`
+- Score: `36` — git repo, 40 TODO/FIXME/WIP hints, plans/backlogs found
+- Evidence: `inventory/evidence/sidecar-v7-v2-520073a262.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/FORENSIK.md`:51 — 22:04–22:38  Autonomous Loop Bug-Fix
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/FORENSIK.md`:271 — **Warum:** `c27535f` "Mega-Session: Alignment Fix + Quality P0/P1" — ein Commit für eine gesamte Session macht Review unmöglich und versteckt Scope-Creep.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/FORENSIK.md`:275 — ### R6: Observer-Peer-Review bei selbstreferenziellem Bau
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/FORENSIK.md`:277 — **Was:** Wenn Claude Sidecar selbst baut, wird nach jedem Feature ein externes Review gemacht (Codex oder MiniMax) mit der Frage: "Löst das das ursprüngliche Problem besser oder baut es das System größer?"
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/HANDOFF.md`:567 — 5. **Safety als "optional aber blockt"** — fail-open bei Load-Fehler. Unveraendert uebernehmen, kein Umbau.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/HANDOFF.md`:627 — """Dispatch ein Event zum registrierten Handler. Fail-open."""
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/BERICHT.md`:39 — SafetyGate wird per `try/except` geladen. Wenn das Laden fehlschlaegt: fail-open. Bei einem Import-Error in `sidecar_samuel` wuerden `rm -rf` und `force-push` NICHT mehr blockiert.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V7-V2/BERICHT.md`:41 — Fail-open ist richtig fuer Coaching-Features. Fail-open ist FALSCH fuer destruktive Gate-Blocker. Safety braucht eine eigene fail-closed Schicht unabhaengig vom Extension-System.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v8-9f03a02b07.md

@@ -0,0 +1,39 @@
+# Sidecar-V8
+
+- Project ID: `sidecar-v8-9f03a02b07`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (7 tracked changes)
+- Branch: `main`
+- Technologies: agent, python
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `60` — git repo, tracked dirty work, 8 TODO/FIXME/WIP hints, README present, agentized
+- Evidence: `inventory/evidence/sidecar-v8-9f03a02b07.json`
+
+## README summary
+
+# Sidecar V8 Implementation repository for `Sidecar V8`. This repo is for source code, tests, and runnable implementation only. Planning does **not** happen here. ## Planning source of truth All planning, architecture, research, and decision documents live in: - `~/Projekte/Sidecar-V8-Plan/` The canonical Claude -> Codex handoff format lives in:
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/AGENTS.md`:17 — - `Codex`: implementation, test execution, review, technical validation
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/AGENTS.md`:42 — - Keep canonical decisions in sync with the plan repo if implementation reveals a plan bug.
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/tests/test_phase_tracker.py`:17 — assert tracker.update_from_prompt("fix this bug") == "debugging"
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/tests/test_phase_tracker.py`:79 — tracker.record_event({"type": "prompt", "content": "fix the login bug now"})
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/tests/test_phase_tracker.py`:80 — tracker.record_event({"type": "prompt", "content": "fix the login bug please"})
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/tests/test_context.py`:104 — state.last_safety_warning = "[SIDECAR-V8] Warning: possible secret exposure detected. Please review."
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/sidecar_core/core.py`:292 — summary="Evolution docs justify one-sentence, fail-open prompt injection.",
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8/core/sidecar_core/phase_tracker.py`:68 — "bug",
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecar-v8-plan-d6f06ac5d8.md

@@ -0,0 +1,38 @@
+# Sidecar-V8-Plan
+
+- Project ID: `sidecar-v8-plan-d6f06ac5d8`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: dirty (1 tracked changes)
+- Branch: `gh-token-furnace-harness`
+- Technologies: docs/notes
+- Triage status: `active`
+- Risk level: `medium`
+- Score: `59` — git repo, tracked dirty work, 40 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/sidecar-v8-plan-d6f06ac5d8.json`
+
+## README summary
+
+# Sidecar V8 Plan Canonical planning repository for `Sidecar V8`. This repo exists to keep V8 planning separate from: - hidden Claude planning paths under `~/.claude/plans/` - personal document dumps under `~/Dokumente/` - V7 implementation work under `~/Projekte/Sidecar-V7/` ## Purpose Use this repository for planning, architecture, research, evaluation, and decision documents only.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/README.md`:72 — | E8 | Knowledge Objects — consolidate or drop? | open |
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/pull_request_template.md`:29 — ## Human Review Needed?
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:19 — - [ ] `READ_ONLY_AUDIT`
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:20 — - [ ] `TEST_GENERATION`
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:21 — - [ ] `SMALL_FIX`
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:22 — - [ ] `REFACTOR_PLAN_ONLY`
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:23 — - [ ] `PR_REVIEW`
+- `/home/smlflg/Projekte/Sidecar-Workspace/Sidecar-V8-Plan/.github/ISSUE_TEMPLATE/agent_job_contract.md`:24 — - [ ] `DOC_REALITY_SYNC`
+
+## Possible next actions
+
+1. Review tracked dirty work and decide: continue, park, or archive. Do not commit without owner approval.
+2. Cluster TODO/FIXME/WIP hints into one bounded next action.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sidecardashboard-10c973f6a7.md

@@ -0,0 +1,31 @@
+# SidecarDashboard
+
+- Project ID: `sidecardashboard-10c973f6a7`
+- Path: `/home/smlflg/Projekte/Sidecar-Workspace/SidecarDashboard`
+- Last modified: `2026-05-04T18:04:15+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: agent, python
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `18` — git repo, agentized
+- Evidence: `inventory/evidence/sidecardashboard-10c973f6a7.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- No safe TODO/FIXME/WIP snippets found.
+
+## Possible next actions
+
+1. Classify purpose from filenames only; ask owner before writing project README.
+2. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/single-role-325160d775.md

@@ -0,0 +1,40 @@
+# single-role
+
+- Project ID: `single-role-325160d775`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `active`
+- Risk level: `low`
+- Score: `36` — 25 TODO/FIXME/WIP hints, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/single-role-325160d775.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:7 — - [ ] `npm run check` passes with no type errors
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:8 — - [ ] `npm run build` completes successfully
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:12 — - [ ] `bash scripts/check-architecture.sh` passes with no violations
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:13 — - [ ] No `fs` or `path` imports in renderer code
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:14 — - [ ] No Electron IPC in service code
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:15 — - [ ] No React imports in services or main process
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:19 — - [ ] Application starts without errors (`npm run dev`)
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-05/solution/single-role/clean-state-checklist.md`:20 — - [ ] Structured log output appears in console at startup
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/skills-3a3285b8ea.md

@@ -0,0 +1,32 @@
+# skills
+
+- Project ID: `skills-3a3285b8ea`
+- Path: `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/litellm_proxy/skills`
+- Last modified: `2026-03-23T15:16:22+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `12` — 3 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/skills-3a3285b8ea.json`
+
+## README summary
+
+# LiteLLM Skills - Database-Backed Skills Storage This module provides database-backed skills storage as an alternative to Anthropic's cloud-based Skills API. It enables using skills with **any LLM provider** (Bedrock, OpenAI, Azure, etc.) by storing skills locally and converting them to tools + system prompt injection. ## Architecture ```mermaid flowchart TB subgraph "Skill Creation" A[User creates skill with ZIP file] --> B{custom_llm_provider?} B -->|anthropic| C[Forward to Anthropic API]
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/litellm_proxy/skills/README.md`:246 — files=[open("data-analysis.zip", "rb")],
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/litellm_proxy/skills/sandbox_executor.py`:101 — with open(local_path, "wb") as f:
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/litellm_proxy/skills/sandbox_executor.py`:243 — with open(tmp_path, "rb") as f:
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/skills-7c8a2825fa.md

@@ -0,0 +1,37 @@
+# skills
+
+- Project ID: `skills-7c8a2825fa`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/skills`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `31` — 40 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/skills-7c8a2825fa.json`
+
+## README summary
+
+# Skills [中文版](./README-CN.md) This directory contains reusable AI agent skills for the Learn Harness Engineering project. Each skill is a self-contained prompt template that can be loaded by AI coding agents (Claude Code, Codex, Cursor, Windsurf, etc.) to perform specialized tasks. ## Available Skills ### harness-creator Production harness engineering skill for AI coding agents. Helps create, assess, and improve agent harness files (AGENTS.md, feature lists, verification workflows, session continuity mechanisms). - **5 reference patterns**: Memory Persistence, Context Engineering, Tool Registry, Multi-Agent Coordination, Lifecycle & Bootstrap - **Templates**: AGENTS.md, feature-list.json, init.sh, progress.md
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:196 — - [ ] Implementation complete
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:197 — - [ ] Verification passed
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:198 — - [ ] Evidence recorded
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:199 — - [ ] Repository restartable
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:335 — 3. **Extraction timing creates race window** — User can start next turn before background extraction completes.
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/SKILL.md`:350 — 违反这些非明显原则会导致 bug：
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/README.md`:160 — - [ ] Python scripts for automated harness generation
+- `/home/smlflg/Projekte/learn-harness-engineering/skills/harness-creator/README.md`:161 — - [ ] HTML viewer for harness assessment results
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/skills-e21ab65783.md

@@ -0,0 +1,31 @@
+# skills
+
+- Project ID: `skills-e21ab65783`
+- Path: `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/anthropic/skills`
+- Last modified: `2026-03-23T15:16:21+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `10` — 2 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/skills-e21ab65783.json`
+
+## README summary
+
+# Anthropic Skills API Integration This module provides comprehensive support for the Anthropic Skills API through LiteLLM. ## Features The Skills API allows you to: - **Create skills**: Define reusable AI capabilities - **List skills**: Browse all available skills - **Get skills**: Retrieve detailed information about a specific skill - **Delete skills**: Remove skills that are no longer needed
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/anthropic/skills/readme.md`:69 — # Get next page if available
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/llms/anthropic/skills/readme.md`:188 — - `next_page` (str, optional): Pagination token for the next page
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/slackalerting-f9ff5ba89b.md

@@ -0,0 +1,30 @@
+# SlackAlerting
+
+- Project ID: `slackalerting-f9ff5ba89b`
+- Path: `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/integrations/SlackAlerting`
+- Last modified: `2026-03-23T15:16:15+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `8` — 1 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/slackalerting-f9ff5ba89b.json`
+
+## README summary
+
+# Slack Alerting on LiteLLM Gateway This folder contains the Slack Alerting integration for LiteLLM Gateway. ## Folder Structure - `slack_alerting.py`: This is the main file that handles sending different types of alerts - `batching_handler.py`: Handles Batching + sending Httpx Post requests to slack. Slack alerts are sent every 10s or when events are greater than X events. Done to ensure litellm has good performance under high traffic - `types.py`: This file contains the AlertType enum which is used to define the different types of alerts that can be sent to Slack. - `utils.py`: This file contains common utils used specifically for slack alerting ## Budget Alert Types
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/.hermes/.uv-cache/archive-v0/9YH33THSokKdr36ExWiQN/litellm/integrations/SlackAlerting/slack_alerting.py`:1347 — API requests will be rejected until either (a) you increase your monthly budget or (b) your monthly usage resets at the beginning of the next calendar month. <br /> <br />
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/smlfg-bd1773137b.md

@@ -0,0 +1,31 @@
+# smlfg
+
+- Project ID: `smlfg-bd1773137b`
+- Path: `/home/smlflg/Projekte/LuckSurface/smlfg`
+- Last modified: `2026-04-21T15:51:03+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `review`
+- Risk level: `low`
+- Score: `10` — 2 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/smlfg-bd1773137b.json`
+
+## README summary
+
+# Samuel Fleig Agentic operator / meta-engineer for AI-coding workflows. I am not positioning as a classical SWE. My strongest work is operating, comparing, coaching, and evaluating agentic coding systems: Hermes, Codex, Claude Code, OpenCode, OpenClaw, Aider, and related workflow layers. ## What I Do - **Build agentic workflows:** context packs, routing, review loops, human-in-the-loop boundaries. - **Coach AI-coding practice:** help people work with agentic tools without drifting into tool chaos. - **Evaluate behavior:** turn fuzzy prompt/workflow impressions into runs, comparisons, failures, and next steps. - **Surface the work:** convert logs, screenshots, decisions, and evals into public proof through Samuel.SYSTEM.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/LuckSurface/smlfg/README.md`:9 — - **Build agentic workflows:** context packs, routing, review loops, human-in-the-loop boundaries.
+- `/home/smlflg/Projekte/LuckSurface/smlfg/README.md`:11 — - **Evaluate behavior:** turn fuzzy prompt/workflow impressions into runs, comparisons, failures, and next steps.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solid-ec3d38ffe8.md

@@ -0,0 +1,35 @@
+# solid
+
+- Project ID: `solid-ec3d38ffe8`
+- Path: `/home/smlflg/Projekte/symphony/elixir/deps/solid`
+- Last modified: `2026-04-28T12:07:20+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: docs/notes
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `18` — 6 TODO/FIXME/WIP hints, README present
+- Evidence: `inventory/evidence/solid-ec3d38ffe8.json`
+
+## README summary
+
+# Solid Solid is an implementation in Elixir of the [Liquid](https://shopify.github.io/liquid/) template language with strict parsing. ## Basic Usage ```elixir iex> template = "My name is {{ user.name }}" iex> {:ok, template} = Solid.parse(template) iex> Solid.render!(template, %{ "user" => %{ "name" => "José" } }) |> to_string "My name is José"
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:3 — ## Bug fixes
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:12 — ## Bug fixes
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:28 — ## Bug fixes
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:34 — * Fix `replace_last` filter bug with duplicate substrings
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:40 — ## Bug fixes
+- `/home/smlflg/Projekte/symphony/elixir/deps/solid/CHANGELOG.md`:53 — ## Bug fixes
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solution-24369842ef.md

@@ -0,0 +1,39 @@
+# solution
+
+- Project ID: `solution-24369842ef`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `active`
+- Risk level: `low`
+- Score: `35` — 40 TODO/FIXME/WIP hints, agentized
+- Evidence: `inventory/evidence/solution-24369842ef.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/claude-progress.md`:40 — **Next session**: No remaining features. Project 06 is complete.
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:7 — - [ ] `npm run check` passes with no type errors
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:8 — - [ ] `npm run build` completes successfully
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:9 — - [ ] No TypeScript warnings about unused variables or imports
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:13 — - [ ] No `fs` or `path` imports in renderer code (`src/renderer/`)
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:14 — - [ ] No Electron IPC in service code (`src/services/`)
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:15 — - [ ] No React imports in services or main process
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-06/solution/clean-state-checklist.md`:16 — - [ ] All IPC channels defined in `src/shared/types.ts`
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Classify purpose from filenames only; ask owner before writing project README.
+3. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+4. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solution-29b8e9044f.md

@@ -0,0 +1,35 @@
+# solution
+
+- Project ID: `solution-29b8e9044f`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-01/solution`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `review`
+- Risk level: `low`
+- Score: `17` — 3 TODO/FIXME/WIP hints, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/solution-29b8e9044f.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-01/solution/claude-progress.md`:27 — **Next session**: Proceed to Project 02 to add import, detail view, and persistence features.
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-01/solution/data/sample-documents/meeting-summary.txt`:16 — 4. Next Steps
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-01/solution/src/renderer/App.tsx`:52 — // In a real app this would open a file dialog.
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solution-5c505f1d56.md

@@ -0,0 +1,40 @@
+# solution
+
+- Project ID: `solution-5c505f1d56`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `active`
+- Risk level: `low`
+- Score: `36` — 24 TODO/FIXME/WIP hints, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/solution-5c505f1d56.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:7 — - [ ] `npm run check` passes with no type errors
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:8 — - [ ] `npm run build` completes successfully
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:12 — - [ ] `bash scripts/check-architecture.sh` passes with no violations
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:13 — - [ ] No `fs` or `path` imports in renderer code
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:14 — - [ ] No Electron IPC in service code
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:15 — - [ ] No React imports in services or main process
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:19 — - [ ] Application starts without errors (`npm run dev`)
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-04/solution/clean-state-checklist.md`:20 — - [ ] Structured log output appears in console at startup
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solution-b44612ff4e.md

@@ -0,0 +1,36 @@
+# solution
+
+- Project ID: `solution-b44612ff4e`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-03/solution`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `20` — 4 TODO/FIXME/WIP hints, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/solution-b44612ff4e.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-03/solution/session-handoff.md`:41 — - The one-feature-at-a-time policy from AGENTS.md was followed strictly: each feature was implemented, verified, and recorded before moving to the next.
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-03/solution/session-handoff.md`:66 — ### Next Steps
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-03/solution/AGENTS.md`:24 — 6. **Only then** move to the next feature.
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-03/solution/data/sample-documents/meeting-summary.txt`:16 — 4. Next Steps
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/solution-e1178ac8ec.md

@@ -0,0 +1,34 @@
+# solution
+
+- Project ID: `solution-e1178ac8ec`
+- Path: `/home/smlflg/Projekte/learn-harness-engineering/projects/project-02/solution`
+- Last modified: `2026-04-28T12:16:56+00:00`
+- Git repo: no
+- Git status: clean/unknown (0 tracked changes)
+- Branch: ``
+- Technologies: agent, node, typescript
+- Triage status: `review`
+- Risk level: `low`
+- Score: `16` — 2 TODO/FIXME/WIP hints, agentized, plans/backlogs found
+- Evidence: `inventory/evidence/solution-e1178ac8ec.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-02/solution/session-handoff.md`:50 — ### Next Steps
+- `/home/smlflg/Projekte/learn-harness-engineering/projects/project-02/solution/data/sample-documents/meeting-summary.txt`:16 — 4. Next Steps
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Read plan/backlog artifacts and extract current unfinished thread.
+3. Classify purpose from filenames only; ask owner before writing project README.
+4. Inspect agent handoff/async state read-only and summarize stale/running/done work.
+5. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.

ProjectIntakeTower/inventory/project_cards/sonnenuhr-0fd23896b3.md

@@ -0,0 +1,37 @@
+# Sonnenuhr
+
+- Project ID: `sonnenuhr-0fd23896b3`
+- Path: `/home/smlflg/Projekte/Sonnenuhr`
+- Last modified: `2026-05-04T17:33:24+00:00`
+- Git repo: yes
+- Git status: clean/unknown (0 tracked changes)
+- Branch: `master`
+- Technologies: python
+- Triage status: `quick_win`
+- Risk level: `low`
+- Score: `22` — git repo, 7 TODO/FIXME/WIP hints
+- Evidence: `inventory/evidence/sonnenuhr-0fd23896b3.json`
+
+## README summary
+
+No safe README summary found.
+
+## TODO/FIXME/WIP hints
+
+- `/home/smlflg/Projekte/Sonnenuhr/config.py`:47 — with _DEFAULT_CONFIG_PATH.open("r", encoding="utf-8") as fh:
+- `/home/smlflg/Projekte/Sonnenuhr/config.py`:52 — with user_config_path.open("r", encoding="utf-8") as fh:
+- `/home/smlflg/Projekte/Sonnenuhr/config.py`:65 — with user_config_path.open("r", encoding="utf-8") as fh:
+- `/home/smlflg/Projekte/Sonnenuhr/sonnenuhr.py`:116 — fd = os.open(LOCK_PATH, os.O_CREAT | os.O_WRONLY)
+- `/home/smlflg/Projekte/Sonnenuhr/sonnenuhr.py`:145 — with open(STATE_TMP, "w", encoding="utf-8") as fh:
+- `/home/smlflg/Projekte/Sonnenuhr/sonnenuhr.py`:843 — # by switching immediately. The lock clears on the next natural transition.
+- `/home/smlflg/Projekte/Sonnenuhr/sonnenuhr.py`:847 — logger.info("Current theme: %s (locked until next transition)", current_theme)
+
+## Possible next actions
+
+1. Cluster TODO/FIXME/WIP hints into one bounded next action.
+2. Classify purpose from filenames only; ask owner before writing project README.
+3. Produce a handoff artifact under ProjectIntakeTower/jobs/done/<job_id>/.
+
+## Safety note
+
+Read-only intake only. Do not modify this project from an intake job unless Samuel explicitly changes the job mode.