Repository navigation
Conversation
Signed-off-by: Srinivasa Consultants <dinakarjs@gmail.com>
Signed-off-by: Srinivasa Consultants <dinakarjs@gmail.com>
Signed-off-by: Srinivasa Consultants <dinakarjs@gmail.com>
Signed-off-by: Srinivasa Consultants <dinakarjs@gmail.com>
Signed-off-by: Srinivasa Consultants <dinakarjs@gmail.com>
Pin the successful current cA2A differential run and expanded Collector evidence. Retain draft status and all synthetic/delivery acceptance limits. Signed-off-by: Srinivasa Dinakar Jainti <dinakarjs@gmail.com> Signed-off-by: Srinivasa Consultants <78673153+dinakarjs@users.noreply.github.com>
|
@imran-siddique — placement and workflow-authorization follow-through updated 1 November 2026 at Please disposition two concrete items:
CODEOWNERS names you, Rajnish and Carlos; the existing implementation and validation evidence are in the description. I will make the selected placement revision and verify it after the owner decision. This is an update to the existing request; no CI-authorization bypass, deployment, new local suite run or independent review is claimed. Draft retained. Prepared with OpenAI Codex. |
|
@imran-siddique Closure decision requested on current head b60e6ca: please select whether the explicitly unmerged revision-bound inventory belongs on /verify/ or in contributor/review documentation. My recommendation remains moving unmerged evidence to contributor/review documentation and reserving /verify/ for accepted/released capabilities. If this contribution should proceed, its three hosted workflows still require maintainer authorization; action_required is not a test failure. I have retained draft status and have not published the site or bypassed workflow approval. Prepared with OpenAI Codex. |
Propose a revision-bound capability evidence inventory that explicitly marks unmerged evidence.
Parked draft:
b60e6ca0b5e6aea33cd88ed91b1c1a172ad8e2f2. The owner must choose /verify/ versus contributor/review documentation. Our recommendation is contributor/review placement for unmerged evidence, with accepted/released capabilities on /verify/.All three returned workflow runs are
action_required; this is an authorization dependency, not a demonstrated test failure. Resume after placement and workflow authorization. No site publication, bypass of workflow approval or capability acceptance is claimed. Routine refreshes are paused.Prepared with OpenAI Codex; source head and existing review history retained.
Historical description and validation snapshots (superseded by current status above)
Evidence refresh - 30 October 2026
Current head b60e6ca, tested tree f96f0985e0a64d3a03132199d38130e7580e76ab. The cA2A row now cites 877ade76 and successful run 41028279638: paired TCK outcomes remain 79 passed, 156 skipped and 30 deselected per run, with zero protected-handler calls. The Telemetry row cites final head 3cf135ad and successful Collector run 40204492721, retaining saturation/loss, in-memory queue and ordering limitations. Both run head SHAs and relevant job logs were checked. Registry #144's unchanged historical evidence is retained.
Three renderer tests, generated freshness, dash and site validation pass: 90 HTML pages, 2000 local links/assets, unique IDs and valid JSON-LD. No public deployment or released capability claim is made. Existing draft-placement and hosted maintainer authorization decisions remain.
Implements the initial, explicitly bounded inventory proposed in #104.
Add a generated capability table to /verify/ with pinned implementation revisions, successful executable CI runs, evidence class, review date and limitations. Seed it with three reviewed draft contributions (agentrust-io/ca2a#266, agentrust-io/trace-registry#144, agentrust-io/agentrust-telemetry#100), each clearly labeled unmerged. This is an inventory of work under review, not a released-product support or certification claim.
The renderer escapes content, constrains GitHub reference fields and rejects missing limitations, duplicate capabilities and mutable revision references. CI checks generated output freshness and rendering regression tests. Existing verifier behavior and normative specifications are unchanged.
Validation: three Node rendering tests passed; generated table freshness, 90-page local link/JSON-LD check, shared header check, dash style, discovery/sitemap generation and discovery metadata checks all passed.
Maintainers should decide whether draft implementation evidence belongs on the public verify page before merge. No deployment requested. AI-assisted; all commits DCO signed.
Hosted CI is awaiting maintainer approval for this first contribution: House style, Reader journey and Browser TDX verifier are
action_required, not failed test results. Local validation above is complete; no CI approval bypass was attempted.