Fix Copilot auth for leak scanner/fixer workflows via PAT pool

[kubaflo]

No description provided.

[github-actions]

🚀 Dogfood this PR with:

⚠️ WARNING: Do not do this without first carefully reviewing the code of this PR to satisfy yourself it is safe.

curl -fsSL https://raw.githubusercontent.com/dotnet/maui/main/eng/scripts/get-maui-pr.sh | bash -s -- 36446

Or

• Run remotely in PowerShell:

iex "& { $(irm https://raw.githubusercontent.com/dotnet/maui/main/eng/scripts/get-maui-pr.ps1) } 36446"

[Copilot]

Pull request overview

This PR fixes authentication failures in the daily-leak-hunter and leak-fixer GitHub Agentic Workflows by switching them to the repository’s established PAT-pool pattern (shared/pat_pool.md + copilot-pat-pool environment) and recompiling their generated .lock.yml artifacts.

Changes:

• Update both workflow source .md files to import shared/pat_pool.md, run in copilot-pat-pool, and override engine.env.COPILOT_GITHUB_TOKEN via case(needs.pat_pool.outputs.pat_number, …).
• Add on.permissions: {} to force generation of the required pre_activation job per pat_pool.README.md known issue guidance.
• Recompile both .lock.yml workflow twins with gh aw v0.80.9, updating pinned action/container versions accordingly.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
.github/workflows/daily-leak-hunter.md	Switches the hunter workflow to PAT-pool auth via shared/pat_pool.md and copilot-pat-pool, including the required on.permissions: {}.
.github/workflows/daily-leak-hunter.lock.yml	Regenerated compiled workflow reflecting PAT-pool wiring and updated gh-aw pinned action/container versions.
.github/workflows/leak-fixer.md	Switches the fixer workflow to PAT-pool auth via shared/pat_pool.md and copilot-pat-pool, including the required on.permissions: {}.
.github/workflows/leak-fixer.lock.yml	Regenerated compiled workflow reflecting PAT-pool wiring and updated gh-aw pinned action/container versions.

[PureWeen]

Adversarial multi-model review — 3 independent reviewers + consensus.

✅ Clean. This PR correctly conforms daily-leak-hunter and leak-fixer to the established shared/pat_pool.md contract, matching the 9 PAT-pool workflows already in production. Verified against source and the compiled locks:

• Token case() override — all 10 slots '0'..'9' → COPILOT_PAT_0..9 with the correct 'NO COPILOT PAT AVAILABLE' default; byte-identical in both workflows and aligned with the pool's import-schema.
• environment: gh-aw-agents → copilot-pat-pool — not a security regression. copilot-pat-pool is the isolated, protected-branches-only environment, applied to every secret-bearing job (agent, detection, pat_pool, pre_activation, safe_outputs). No cross-environment secret becomes reachable.
• on.permissions: {} — the documented workaround (gh-aw#30790) that forces the no-op pre_activation job the pat_pool job depends on. It does not clobber the real per-job permissions: (agent stays read-only; safe_outputs keeps issues: write).
• Lock/source integrity — recompiling both with gh aw v0.80.9 produces zero diff; both compile with 0 errors / 0 warnings. No stale gh-aw-agents references remain.
• Secret handling — the resolved PAT is excluded from the agent sandbox (--exclude-env COPILOT_GITHUB_TOKEN), all COPILOT_PAT_* are registered for log redaction, and the pat_pool job emits only the selected index (0–9), never a secret.

ℹ️ Non-blocking FYI (pre-existing, repo-wide — not introduced by this PR): In the generated locks, the detection (threat-detection) job's COPILOT_GITHUB_TOKEN references needs.pat_pool.outputs.pat_number, but pat_pool is not in that job's needs: (only activation, agent). Since Actions exposes only direct-dependency outputs, that resolves to the 'NO COPILOT PAT AVAILABLE' literal, so the threat scan fails open. Impact is contained: the detection execute + conclude steps are continue-on-error: true, so detection.result stays success and safe_outputs (if: needs.detection.result == 'success') still runs — issues/PRs are still posted. This is a property of the gh-aw compiler + shared pat_pool.md import (which wires pat_pool into activation/agent needs only) and is identical across all 11 PAT-pool workflows here (9 already in production). It is not fixable in these two source files — raising it purely as something worth a follow-up upstream with gh-aw if reliable threat-detection auth is wanted for pooled-PAT workflows.

---

Test coverage: These are CI/agentic-workflow config files; the compiler (gh aw compile, clean) + schema validation are the applicable checks — both pass.
Prior reviews: Existing approval (@PureWeen) and the Copilot reviewer's no-comment overview; this review is additive (lock/source-sync verification, environment-boundary analysis, and the detection fail-open note were not previously covered).
Methodology: 3 independent reviewers with adversarial consensus. One reviewer initially flagged the detection-token behavior as blocking; on verification its blocking-impact rationale was disproven (fail-open via continue-on-error), so it is reported as a non-blocking, pre-existing note rather than a defect in this PR.