Repository navigation
Description
Decision (maintainer, 2026-09-12)
Personas post as don-petry, not donpetry-bot. This reverses the identity change made by
#1650 / PR #1694 and restores the behaviour ADR-0005 already describes.
Why this is a restoration, not a new direction
ADR-0005 (accepted) already sanctions it. Its §on write-identity says:
"'Write runtime identity' here means code/branch write capability (commit, push, label, close),
not the account a persona posts advisory comments as: an advisory-only persona may still run
under a real account —security-lead, for instance, posts as the owner accountdon-petry."
#1650 changed eight manifests to donpetry-bot without superseding ADR-0005, so the accepted
record and the code have contradicted each other since 2026-09-08. dev-lead was never switched and
still posts as don-petry — so today the fleet is also internally inconsistent.
It also fixes a live regression. #1734: DON_PETRY_BOT_GH_PAT authenticates as donpetry-bot
but lacks permission to write issue comments here, so the post step 403s and the advisory is
discarded — not persisted anywhere. A real qa-lead invocation on epic #1723 was lost this way,
twelve minutes after #1694 merged. GH_PAT_DON_PETRY is known-good: dev-lead posts with it
continuously today.
The trade-off this accepts — state it, do not bury it
#1650's rationale was real and is now consciously traded away. persona-mention-reusable.yml's
pre-filter carries two independent recursion axes:
github.event.comment.user.login != 'donpetry-bot' && # actor axis
github.event.comment.user.login != 'github-actions[bot]' &&
!contains(github.event.comment.body, '<!-- persona:') && # marker axisWith personas posting as don-petry, the actor axis no longer covers persona output — a human
account cannot be added to that exclusion list without making the persona un-invocable by the very
maintainer who invokes it most. So the marker axis becomes the sole mechanical guard.
That is acceptable, for reasons worth recording:
- The marker is workflow-enforced, not prompt-enforced.
persona-runner-reusable.ymlapplies
pr_ensure_markerin the post step — "the only writer" — so an agent that forgets the marker
cannot post an unmarked comment, because it cannot post at all. Prompt-only marker enforcement is
precisely how feat: implement issue #846 — [Phase 2] Cache the LSP index and enforce the 30s cold-start SLA with auto-skip #860 produced 1,481 acks in 4.5h; this design deliberately does not rely on it. - It has been proven live. The [persona-runtime soak] qa-lead mention smoke test #1300 soak posted a marked advisory, which triggered a router run
that returnedskipped— the guard caught its own marker. - A third, weaker protection remains:
prompts/<id>/advisory.mdforbids writing a literal
@petry-projects/<role>in the body, so an advisory should not contain a routable handle at all.
The residual risk is a single-axis guard. If the marker enforcement in the post step is ever
weakened or bypassed, nothing else stops a persona comment from re-routing. That makes
pr_ensure_marker and its bats coverage load-bearing in a way they were not before — AC #6 below
makes that explicit rather than tribal.
Acceptance Criteria
- ADR-0008 is authored at
docs/architecture/adr/0008-personas-post-as-the-owner-account.md,
following ADR-0000's process: Nygard's five sections in order (Title/Status/Context/
Decision/Consequences),Status: accepted. Context names [qa-lead S7] Stop posting persona advisories under a human maintainer's PAT #1650's experiment, ADR-0005's
existing position, and persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734's regression. Consequences states the single-axis trade-off above
honestly, including the residual risk — an ADR that records only the upside is not a decision
record. - ADR-0005 is NOT edited. ADR-0000 makes an accepted ADR immutable, with the single permitted
edit being asuperseded-bystatus change. ADR-0008 aligns with ADR-0005 rather than reversing
it, so no supersede link is warranted — ADR-0008's Context should say so explicitly, so a future
reader does not assume an unrecorded conflict. - The eight manifests are updated —
business-analyst,devops-lead,pr-review,qa-lead,
scrum-master,security-lead,solution-architect,sre-lead— to
account: don-petry/credential: GH_PAT_DON_PETRY.dev-leadis already correct and must not
be touched.validate-personas.py'scheck_identity(theGH_PAT_<ACCOUNT>schema rule) must
still pass. tests/resolve_persona_identity.batsis updated, including the two tests that assert the
bot identity by name ("qa-lead posts as the bot identity donpetry-bot, not a human maintainer"
and its credential sibling). Keep the generic tests that prove identity is read from the manifest
and never defaulted — those encode fix(dev-lead): resolve identity from persona manifest, not shared vars.BOT_USER (#1316) #1317's guarantee and are still correct.- The stale rationale in
persona-runner-reusable.ymlis corrected. Its header currently asserts
the persona posts as a bot "never a human maintainer — a human byline halves the recursion
guard". That comment will be false after this change; replace it with the ADR-0008 reasoning and
cite the ADR. The credential-resolution logic already accepts both secrets and needs no behaviour
change. - The marker guarantee is reinforced, since it is now load-bearing. Add a comment at the
pr_ensure_markercall site in the post step recording that it is the sole mechanical recursion
guard for persona output under ADR-0008, and confirmtests/persona_runner.batsstill covers the
forgot-the-marker path (pr_ensure_marker PREPENDS the marker when the agent forgot it). - persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734 is verified fixed by a live post, not by inspection. After the change, invoke a persona
and confirm an advisory actually lands — authordon-petry, marker present as the first body
line — and link the run. This is the acceptance bar, and it is the step that was skipped last
time: [qa-lead S7] Stop posting persona advisories under a human maintainer's PAT #1650's own AC feat: feature ideation framework — adversarial proposer/challenger/synthesizer #6 required exactly this live verification, it was never performed, and the
first real invocation after merge 403'd and lost its advisory.
Tasks / Subtasks
- Author ADR-0008 per ADR-0000's template and immutability rules. (AC: test issue from agent #1, Go-live improvements for PR review agent #2)
- Flip the eight manifests; leave
dev-leadalone. (AC: Add @claude delegation, auto-merge, and rebase handling #3) - Update the identity bats expectations. (AC: Optimize review: small-PR and incremental fast paths #4)
- Correct the runner's header rationale and cite ADR-0008. (AC: feat: add Copilot engine support via REVIEW_ENGINE toggle #5)
- Document and re-confirm the marker guard's now-sole status. (AC: feat: feature ideation framework — adversarial proposer/challenger/synthesizer #6)
- Live-verify a posted advisory and link the run; close persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734. (AC: Workflow should skip no-op PRs in MAX_PRS count #7)
Dev Notes
- Do NOT change the router's actor-exclusion list in
petry-projects/.github
(persona-mention-reusable.ymllines ~90-91). Addingdon-petrythere would make personas
un-invocable by the maintainer who invokes them most — the exclusion exists to stop bot output
re-routing, and a human account cannot serve both roles.donpetry-botshould remain excluded: it
is stilldev-lead's review identity and other automation posts as it. GH_PAT_DON_PETRYis known-good —dev-leadposts with it continuously. No new secret needs
provisioning, which is the main reason this resolves persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734 immediately rather than requiring a
permissions change on the bot account.- Do not "simplify" by removing
DON_PETRY_BOT_GH_PATfrom the reusable'ssecrets:block — other
paths still reference the bot credential, and the resolution logic accepting both is what makes
this change a one-line manifest edit per persona. - Approval mechanics are unaffected:
require_last_push_approvalcompares the approver to the last
pusher. Personas post comments; they do not push.donpetry-botremains the PR-review approver.
Project Structure Notes
New docs/architecture/adr/0008-*.md; eight personas/*/persona.yml; tests/resolve_persona_identity.bats;
comment-only edits in .github/workflows/persona-runner-reusable.yml. No change in petry-projects/.github.
References
- ADR-0005 (
docs/architecture/adr/0005-advisory-by-default-personas.md) — already permits a real account - ADR-0000 — ADR process, immutability, numbering
- [qa-lead S7] Stop posting persona advisories under a human maintainer's PAT #1650 / PR feat: implement issue #1650 — [qa-lead S7] Stop posting persona advisories under a human maintainer's PAT #1694 — the change being reversed; its unperformed AC feat: feature ideation framework — adversarial proposer/challenger/synthesizer #6 is this issue's AC Workflow should skip no-op PRs in MAX_PRS count #7
- persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734 — the live regression this fixes
- fix(dev-lead): resolve identity from persona manifest, not shared vars.BOT_USER (#1316) #1317 — identity resolved from the manifest, never a shared default (preserve)
docs/postmortems/2026-06-pr-860-runaway.md— why marker enforcement lives in the workflow- petry-projects/.github
standards/persona-standards.md§5.1 — mechanism is identity-agnostic; no change needed
Activity
don-petry commentedon Sep 12, 2026
Dev-Lead Implementation Plan
Issue: #1765 — Personas post as don-petry, not donpetry-bot — record as ADR-0008 and apply across the fleet (fixes #1734, reverses #1650)
Scope
Author ADR-0008 recording that advisory personas post as the owner account
don-petry, flip the advisory persona manifests todon-petry/GH_PAT_DON_PETRY, update the identity bats, and correct the runner's now-stale rationale — reinforcing the marker guard that becomes the sole mechanical recursion guard.pr-reviewis intentionally NOT flippedAC #3 lists eight manifests including
pr-review. I am flipping only the seven advisory personas and leavingpr-reviewasdonpetry-bot, because flippingpr-reviewcontradicts the rest of this issue and would break the merge gate:donpetry-botremains the PR-review approver" and "it is stilldev-lead's review identity."donpetry-bot… They are deliberately different accounts."pr-review.ymlresolvesBOT_USERfrompersonas/pr-review/persona.ymlruntime.identity.account— the account it approves PRs as.dev-leadauthors PRs asdon-petry; GitHub forbids approving your own PR, sopr-reviewapproving asdon-petrybreaksrequire_last_push_approval/CODEOWNERS.GH_PAT_DON_PETRYmay also be fine-grained, whichpr-review.ymlrejects foraddPullRequestReview.pr-review's persona identity governs review/approval, which is out of that scope.The seven flipped:
business-analyst,devops-lead,qa-lead,scrum-master,security-lead,solution-architect,sre-lead.dev-leadalready correct;pr-reviewleft as bot. Please confirm if pr-review was genuinely intended to flip.Implementation Checklist
docs/architecture/adr/0008-personas-post-as-the-owner-account.md(Nygard 5 sections,accepted; names [qa-lead S7] Stop posting persona advisories under a human maintainer's PAT #1650, ADR-0005 alignment, persona-runner: qa-lead advisory lost — DON_PETRY_BOT_GH_PAT gets HTTP 403 posting issue comments (regression from #1694) #1734; honest single-axis trade-off + residual risk) (AC test issue from agent #1, Go-live improvements for PR review agent #2)account: don-petry/credential: GH_PAT_DON_PETRY(AC Add @claude delegation, auto-merge, and rebase handling #3, minus pr-review).github/workflows/persona-runner-reusable.yml+ cite ADR-0008 (AC feat: add Copilot engine support via REVIEW_ENGINE toggle #5)pr_ensure_markerpost-step call site recording its now-sole mechanical-guard status (AC feat: feature ideation framework — adversarial proposer/challenger/synthesizer #6)Tests to Write
tests/resolve_persona_identity.bats— the two qa-lead tests to assertdon-petry/GH_PAT_DON_PETRY; keep the generic manifest-driven tests (fix(dev-lead): resolve identity from persona manifest, not shared vars.BOT_USER (#1316) #1317)tests/persona_runner.batsstill covers the forgot-the-marker path (AC feat: feature ideation framework — adversarial proposer/challenger/synthesizer #6)Test Command
scripts/run-bats.sh tests/resolve_persona_identity.bats tests/persona_runner.bats+python3 personas/validate-personas.py+bash .dev-lead/scripts/dev-lead-lint.shNote on AC #7 (live verification)
A live persona invocation is a post-merge step — the dev-lead workspace does not push. I will flag it for post-merge execution in the final report; it cannot be performed from this pre-push workspace.
don-petry commentedon Sep 12, 2026
Dev-Lead: Implementation Complete — PR #1768
The implementation for issue #1765 is durable: commits are pushed and a pull request is open. This record is backed by verifiable artifacts you can confirm in one click.
7be6a2132d8b40f7f7d4286980458a4c4e5f15afReview happens on the PR. Acceptance criteria and test results are evidenced by the PR's diff and CI checks — not asserted here. If a later stage fails, this claim is retracted in place.
feat: implement issue #1765 — Personas post as don-petry, not donpetr…