journal-evidence-export
Skill22 starsChanged 3 months agoApache-2.0View on GitHub
Fetch run evidence from the manzanasd journal - list runs, page entries, download screenshot artifacts, upload your own evidence, and export a PR-ready markdown summary. Use when a task needs proof of what was done on a simulator (QA evidence for a PR, debugging a run).
- Sends data out
In this repository8 other files it gives its agents.
- broker-federationSkill
- build-and-testSkill
- drive-a-mirrored-iphoneSkill
- lease-and-drive-a-simSkill
- park-thaw-semanticsSkill
- run-evalsSkill
- run-the-daemonSkill
- ui-audit-and-predicatesSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Every mutating op under a lease is journaled (append-only JSONL + content- addressed artifacts).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Journal evidence export
Every mutating op under a lease is journaled (append-only JSONL + content-
addressed artifacts). The run ID is the lease ID, so you always know your
run without extra bookkeeping. Journal-disabled daemons (--journal-dir "")
answer 501 on all of these.
The one-call PR evidence export
manzanas journal export $LID -o evidence.md # or: --format json
curl -s $D/v0/journal/$LID/export.md > evidence.md # same markdown, raw HTTPReturns a PR-comment-ready markdown summary: run metadata (agent, purpose,
target, runtime), an ordered action table with failures highlighted, and
the artifact list. Paste it into the PR body or a comment. Works after the
lease is released. --journal-dir ~/.manzanasd/journal exports offline
straight from the on-disk store (daemon stopped, or a copied run dir).
MCP agents have the same export as the journal_export tool; the GitHub
Action posts it automatically with journal-evidence: "true".
Inspect a run
curl -s $D/v0/journal | jq '.runs[0]' # newest runs first
curl -s "$D/v0/journal/$LID?limit=100" | jq '.entries[] | .payload.action'
# next_seq != 0 → pass it back as from_seq for the next pageEntry kinds: lease, action, observation, state, segment. HID
actions record ax_before/ax_after tree hashes when dispatched with
payload "ax_hashes": true — cheap before/after UI evidence.
Fetch artifacts (screenshots)
A successful screenshot action is stored automatically; entries carry
artifacts: [{path, sha256, bytes}]:
curl -s "$D/v0/journal/$LID?limit=1000" \
| jq -r '.entries[].payload.artifacts[]?.path'
curl -s -o shot.png "$D/v0/journal/$LID/artifacts/artifacts/<sha256>.png"Tip: dispatch screenshots with "inline": false to keep responses small and
pull the bytes from the journal afterwards.
Upload your own evidence
While the lease is still active (immutable after release/expiry — upload before you release):
manzanas journal upload $LID shot1.png shot2.png # --kind observation|screenshot|video
# or raw HTTP:
curl -s -X POST "$D/v0/journal/$LID/artifacts?name=final-state&kind=screenshot" \
--data-binary @shot.pngkind: observation (default over HTTP; the CLI defaults to
screenshot), screenshot, or video.
Durability rule: evidence cited in a GitHub issue, PR, or FIXED/closing comment must be uploaded as a journal artifact (cite the run id + artifact name) or attached directly to the issue — bare session-local filenames die with the session box and are not acceptable evidence (see docs/agent-qa.md, "QA evidence durability").
Live tail
manzanas --daemon $D journal tail $LID --follow (WS journal.tail replays
from from_seq then streams new entries).
Retention
Runs are GC'd by --journal-max-age (default 7 days) and
--journal-max-bytes (default 2 GiB); live-lease runs are never reclaimed.
Export evidence promptly — don't rely on the journal as long-term storage.
For evidence that must outlive retention, also attach the images to the
GitHub issue/PR.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

