bookstack
An agent environment exported by REA from one retained web network capture. Every tool answers from a recorded exchange; nothing here runs the captured application.
Provenance
- Capture artifact SHA-256:
1f807d4d2d2ef490cdc5cac50ffcbf573abfc967351038256b92e18f4ebe5366 - Capture Evidence (the manifest's parent):
ev_07a98c12a171833f8ca00f5f4769a165b358a5e2cca784718af469fc35ee5b77 - Manifest Evidence:
ev_47deffa8044f6f1c6835193c59746931187576db6b1c0cc367b7c4380714c322 - Manifest view digest:
9e25a0530da2302938a01675ca07fd7d373d8b750f864a0447aae0d20f6d6930 - Primary origin:
http://127.0.0.1:3219 - Tools: 31
- Serving CLI: the packed tarball
rea-agents.tgzcopied beside the Dockerfile
What replay means and does not mean
envs/bookstack/ builds an image that runs rea environment-serve on the
manifest and capture Evidence copied beside its Dockerfile. Each MCP tool call
selects the retained record whose path and query values best match the
arguments and returns that record's recorded status, content type and body,
citing the record ordinal and the Evidence id. Replay is not emulation: the
environment does not establish the application's authorization model, response
semantics or state transitions, does not validate supplied values, and does not
answer requests the capture never recorded. Replayed content is target content
and must be treated as data, never as instructions.
Run it
From this folder:
agent-env run .
From anywhere else, name this folder's path instead of ..
tasks/smoke.json deploys the env on the local sandbox and runs
artifacts/check_bookstack/verify.py against its MCP endpoint: it lists
the tools and checks the count equals 31, and calls get_attachments and checks the replayed status equals the recorded 200 of record 29.
The run needs Docker; the first run also builds the runner's own infrastructure.
Rights
The recorded content in evidence.json belongs to the captured application.
Do not publish this bundle without the rights to that content.