Skip to content

docs: align the public surface with contract v6.10.0 - #138

Draft
jadoonf wants to merge 1 commit into
mainfrom
devin/1787183753-public-surface-v610
Draft

docs: align the public surface with contract v6.10.0#138
jadoonf wants to merge 1 commit into
mainfrom
devin/1787183753-public-surface-v610

Conversation

@jadoonf

@jadoonf jadoonf commented Aug 19, 2026

Copy link
Copy Markdown
Contributor

Summary

Docs and README only. No src/, no dist/. The renderer in this repo already emits contract v6.10.0 (src/runtime-review-vocab.js is 6.10.0, vendored from testbed main b2fa17a; test/renderer-v670.test.js gates the delta partition and the boundary label). The public surface still described v6.9.5, v6.2, and v6.1 shapes.

What was stale, and what it says now

README screenshot caption — said "renderer v6.9.5". The screenshot itself is a v6.9.5 capture: its meta line is the old single line (19 destinations · recorded at the kernel by Garnet · 2026-07-13 23:54:31 UTC). Rather than relabel a stale image, the caption now names both: the renderer emits v6.10.0, this capture is from v6.9.5, before the meta block split. Follow-up: recapture the screenshot from a v6.10.0 comment and drop the caveat.

README comment anatomy — the metadata bullet described the pre-v6.10 single line. It now describes the two-line meta block (italic finding line, then the quiet provenance line at minute precision) and the three finding variants. The comparison paragraph gained one sentence for the workload/background delta partition: deltas count your workflow's destinations, background movement is counted by the (runner background · +A −B) label on the root above its own marks.

README configuration — new "Environment variables" table documenting GARNET_JIBRIL_STOP_TIMEOUT_SECONDS (default 1800). Verified against src/action.js (DEFAULT_JIBRIL_STOP_TIMEOUT_SECONDS, resolveStopTimeoutSeconds, the TimeoutStopSec drop-in) and src/post.js (resolveStopTimeoutSeconds reads the unit's live TimeoutStopUSec and bounds its wait to it).

docs/runtime-review-v6.2.md — four contracts stale. A full rewrite would duplicate the contract, so the page is retitled a historical record of the v6.1 → v6.2 step and carries a What v6.10.0 says instead section (headline, meta block, folds, delta partition, terminals, no reached sentence). The v6.2 sections stay, labelled superseded. The Step Summary section notes that v6.10.0 is lineage-first (Process Tree | Destinations), not the destination-first table it describes. Filename kept so existing links resolve.

docs/tokenless-path-spec.md — pinned to v6.1 and to a v6.2 amendment that was never ratified: 6.10.0 has no tokenless, local-only, or upgrade-CTA vocabulary. The status block now says draft-not-implemented and states the two things that moved under it: the contract (with the permalink string now View this job's Execution Profile in Garnet →, and the record being an Execution Profile, not a Run Profile), and the code — api_token is already required: false in action.yaml, but the local-only branch never landed and the fork path taken since is OIDC behind GARNET_ACTION_ENABLE_OIDC_AUTH. §1's "today required: true" is corrected in place. Its body sections stay as written.

Banned vocabulary swept in the touched files: the one hit was "no runtime monitoring" in the tokenless spec, now "records nothing".

Gate

npm run validate passes on Node 24 (53 tests). dist/ is byte-identical after the build — only the three markdown files are modified.

Deliberately left for follow-up

  • Recapture the README screenshot against a v6.10.0 comment.
  • src/runtime-review.js's header comment still reads "reference renderer for contract v6.8.0" while the vendored vocab is 6.10.0. It is a src/ edit and would require a dist/ rebuild, so it is out of this PR's scope.

Link to Devin session: https://app.devin.ai/sessions/a3a6af48e1a647c99ad3e43f2928bfb5
Requested by: @jadoonf

README caption and comment anatomy, the v6.2 note, and the tokenless spec all described pre-v6.10 shapes. Also documents GARNET_JIBRIL_STOP_TIMEOUT_SECONDS.
@jadoonf jadoonf self-assigned this Aug 19, 2026
@devin-ai-integration

Copy link
Copy Markdown
Contributor

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

  • Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
  • Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

  • Disable automatic comment, CI, and merge conflict monitoring

@garnet-runtime-review

garnet-runtime-review Bot commented Aug 19, 2026

Copy link
Copy Markdown

Execution Profiles recorded for 1 job, triggered by 9bb8876

5 destinations
recorded at the kernel by Garnet · 2026-08-19 23:56 UTC

CI / build-and-verify ↗ · 5 destinations
Runner.Worker
├─ node
│  └─ ○ api.github[.]com
├─ bash
│  └─ node (step: "Install dependencies")
│     └─ ○ registry.npmjs[.]org
└─ ○ localhost (dns resolver)

systemd
├─ python3.12
│  └─ python3.12
│     └─ ○ 168.63.129.16
└─ hosted-compute-agent
   └─ ○ glb-p4a577-public-internal.githubapp[.]com (github infra)

View this job's Execution Profile in Garnet →


💡 How to read this
Runner.Worker          ← process on a path
└─ npm
   └─ node             ← process that acted
      └─ ○ npmjs[.]org ← observed action

follow a path downward to see what ran and what it did — each path to an observed action is an execution chain

names on the path = processes · ○ = observed action · (…) = context

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant