fix(chat): restrict grounding to public summaries

[JSONbored]

Motivation

• Prevent contributor-visible @gittensory chat prompts from including private agent decision-pack details (e.g. why/blockedBy) that could leak confidential readiness/credibility signals when a PR author is authorized to run chat.
• Keep chat Q&A grounded in deterministic, public-safe facts while preserving the existing publicSafeSummary path for useful contributor guidance.

Description

• Limit the chat grounding bundle in compactChatSignalBundle to include only actionType, status, and a redacted publicSafeSummary, removing the raw why and blockedBy arrays from the serialized prompt payload in src/services/ai-chat-qa.ts.
• Update the ChatGroundingAction/bundle types and the prompt construction so the model never receives why/blockedBy fields.
• Add regression unit tests in test/unit/ai-chat-qa.test.ts asserting that private rationale/blocker examples are omitted from the compacted grounding and that provider prompts no longer serialize "why" or "blockedBy".

Testing

• Ran the targeted unit suite with npm test -- --run test/unit/ai-chat-qa.test.ts, and all tests in that file passed.
• Ran static type checks with npm run typecheck, which passed without errors.
• Ran a scoped coverage invocation npm run test:coverage -- --run test/unit/ai-chat-qa.test.ts; the targeted tests passed but the repository-wide coverage threshold caused the coverage command to fail due to global thresholds (this is expected when running a single-file scope in this repository).

---

Codex Task

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.20%. Comparing base (e04e847) to head (64869e5).
⚠️ Report is 3 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5106      +/-   ##
==========================================
- Coverage   94.20%   94.20%   -0.01%     
==========================================
  Files         470      470              
  Lines       39736    39734       -2     
  Branches    14501    14501              
==========================================
- Hits        37435    37433       -2     
  Misses       1645     1645              
  Partials      656      656              

Flag	Coverage Δ
shard-1	46.40% <ø> (-0.02%)	⬇️
shard-2	33.64% <ø> (-0.22%)	⬇️
shard-3	31.36% <ø> (-0.80%)	⬇️
shard-4	32.77% <ø> (+0.96%)	⬆️
shard-5	33.44% <ø> (-0.10%)	⬇️
shard-6	45.10% <ø> (+0.20%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/services/ai-chat-qa.ts	98.70% <ø> (-0.04%)	⬇️

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[gittensory-orb]

Warning

🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨

⏸️ Gittensory review result - manual review recommended

_{Review updated: 2026-07-11 17:27:12 UTC}

2 files · 1 AI reviewer · 2 blockers · readiness 93/100 · CI green · unstable

⏸️ Suggested Action - Manual Review

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.
• Maintainer requires a linked issue — Link the relevant issue (for example Closes #123) before opening the PR.

Review summary
This narrows the chat-Q&A grounding bundle to actionType/status/redacted publicSafeSummary, dropping raw why/blockedBy arrays that could leak private decision-pack rationale (blocker codes, credibility signals) into a contributor-visible prompt. The change is mechanical and consistent: the ChatGroundingAction type, compactChatSignalBundle, and buildChatPrompt's JSON.stringify output all agree on the narrower shape, and the updated tests assert both structural equality of the compacted bundle and absence of "why"/"blockedBy" keys in the serialized prompt sent to the provider. redactGroundingText and its private-term regexes are untouched, so existing redaction of anything still surfaced (objective, summary, publicSafeSummary, freshnessWarnings) is preserved.

Nits — 5 non-blocking

• The PR description doesn't cite a linked issue number (only references prior PR feat(commands): add @gittensory chat <question> — grounded LLM Q&A via local Ollama #4595 in comments) — confirm this closes a tracked issue per repo convention.
• test/unit/ai-chat-qa.test.ts:bundleFixture still builds action.why/blockedBy fixture data even though they're now always stripped from output — fine for now since AgentRunBundle's action type still carries them, but worth a comment noting they're intentionally retained only as compaction-input fixtures.
• Consider a short code comment at the ChatGroundingAction type definition (src/services/ai-chat-qa.ts:66) explicitly stating why/blockedBy are excluded by design, so a future edit doesn't accidentally re-add them without re-reading the redaction rationale in the nearby comment block.
• PR description: nit: please add an explicit eligible issue link or closing reference for this change; the diff references `feat(commands): add @gittensory chat <question> — grounded LLM Q&A via local Ollama #4595` in existing comments/tests, but the PR body itself only links a Codex task.
• src/services/ai-chat-qa.ts:186: nit: `publicSafeSummary` is still included even when redaction trims it to an empty string, so consider filtering empty summaries or falling back to a generic public-safe action label if empty summaries are possible.

Concerns raised — review before merging

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.
• Maintainer requires a linked issue — Link the relevant issue (for example Closes #123) before opening the PR.

Signal	Result	Evidence
Code review	❌ 2 blockers	1 reviewer
Linked issue	⚠️ Missing	No linked issue or no-issue rationale found.
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (no linked issue context).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 45 registered-repo PR(s), 37 merged, 414 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	JSONbored; Gittensor profile; 45 PR(s), 414 issue(s).
Gate result	❌ Blocking	Repo-configured hard blocker found.
Improvement	✅ Minor	risk: clean · value: minor — Code changes are accompanied by test evidence.

Review context

• Author: JSONbored
• Role context: owner (maintainer lane)
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: not available
• Official Gittensor activity: 45 PR(s), 414 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Treat this as maintainer-lane context rather than normal contributor-lane activity.
• Explain no-issue PR.
• Link the issue being solved, or explicitly explain why this is a no-issue PR.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

[BETA] Chat with Gittensory

Ask Gittensory a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @gittensory ask <question> answers contribution-quality Q&A with source citations and freshness.
• @gittensory chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @gittensory mention with a real question is routed to the closest matching read-only command automatically -- no exact syntax required.

Full command reference: https://gittensory.aethereal.dev/docs/gittensory-commands

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

• Re-run Gittensory review