feat(tooling): pure issue-drafting core + CLI to expand a loose prompt into a gate-ready draft

[JSONbored]

Summary

• Adds the pure, adapter-agnostic issue-drafting core (src/services/issue-drafting.ts) required by the issue's ⚠️ shape mandate: no process.argv, no fs, no IO — the searchable corpus is caller-supplied data, so a future ORB dashboard API route (a Worker with no filesystem) can be a second thin consumer of the identical entrypoint. Mirrors the signal-tracking.ts / signal-tracking-wire.ts pure-core/host-adapter split.
• Grounding pipeline: tiered term extraction from the loose prompt (backticked exact > path > identifier > plain word, each tier consumed before the next scans so fragments never double-extract), case-insensitive corpus search with definition-lines-first / live-code-above-tests ranking, and word-tier terms restricted to files whose path contains them — plain vocabulary can never ground to random comment prose.
• Draft assembly emits the heavy template (Context / Requirements / Deliverables / Test Coverage Requirements / Expected Outcome / Links & Resources) with real path:line citations, one Requirements bullet per anchor file (not per term), the > ⚠️ Required pattern callout, an explicit ⚠️ UNGROUNDED marker at the exact spot any specific term found no precedent, and <!-- MAINTAINER: ... --> markers on every section a human must fill — it never invents an unverified requirement.
• scripts/draft-issue.ts is the first thin consumer: reads the prompt (--prompt / --prompt-file), walks the checkout for the corpus, writes the draft to a local file. Never publishes; labels/milestone/relationships stay maintainer decisions (the issue's hard boundary).

Closes #8103

Performance note (the issue's own acceptance bar)

Ran against one-line prompts approximating two real, recently-shipped sub-issues of #8082:

• calibration: compute REGRESSED-verdict track-record statistics from ORB's persisted backtest results #8140-shaped prompt ("aggregate REGRESSED-verdict track-record statistics from the persisted backtest comparisons via computeRegressedVerdictTrackRecord, printed by a read-only CLI mirroring scripts/backtest-corpus-export.ts"): top citations were packages/loopover-engine/src/calibration/backtest-track-record.ts:27 (the exact function definition), the real track-record CLI, and backtest-corpus-export.ts's usage lines — the same anchors the real calibration: compute REGRESSED-verdict track-record statistics from ORB's persisted backtest results #8140 issue body cited. 0 UNGROUNDED markers.
• Extend the backtest check to logic/regex changes via a dedicated CI job, using captured raw context #8139-shaped prompt: found scoreBacktest's definition, backtest-threshold.ts's own "logic/regex-change backtesting needs raw context" comment line, and the linked_issue_scope_mismatch rule-id constant — again the anchors the real spec used.

Honest limits: it grounds and cites; it does not decompose scope or write Deliverables — those stay explicit MAINTAINER markers. Word-tier matches can still surface plausible-but-tangential anchors (e.g. "verdict" → gate-verdict-calibration.ts); they're real files with definition lines, and the maintainer edit pass is the filter. A deliberate non-goal: no AI/model call anywhere — output is deterministic (same prompt + corpus ⇒ byte-identical draft, tested).

Scope

• The PR title follows type(scope): short summary Conventional Commit format, for example fix(api): restore profile access checks.
• This PR is focused and does not mix unrelated backend, UI, MCP, docs, dependency, and deploy changes.
• This follows CONTRIBUTING.md and does not reintroduce GitHub Pages, VitePress, site/, or CNAME.
• I linked a currently open issue this PR resolves (e.g. Closes #123) — a linked open issue is required for every contributor PR.

Validation

• git diff --check
• npm run actionlint
• npm run typecheck
• npm run test:coverage locally; codecov/patch requires ≥99% coverage of the lines AND branches you changed (aim for 100% on your diff so CI variance does not fail near the threshold). Global coverage is a non-blocking trend with a loose 90% backstop, not the gate.
• npm run test:workers
• npm run build:mcp
• npm run test:mcp-pack
• npm run ui:openapi:check
• npm run ui:lint
• npm run ui:typecheck
• npm run ui:build
• npm audit --audit-level=moderate
• New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries

If any required check was skipped, explain why:

• Full npm run test:ci ran green end-to-end (includes every box above). The core has 100% line AND branch coverage (24 tests, scoped v8 report: 58/58 branches). scripts/draft-issue.ts is thin IO glue (arg parsing, checkout walk, file write) under the established export-d1-data.ts exemption.

Safety

• No secrets, wallet details, hotkeys, coldkeys, user PATs, private keys, raw trust scores, private rankings, or private maintainer evidence are exposed.
• Public GitHub text stays sanitized, low-noise, and does not imply compensation guarantees or optimization tactics.
• Auth, cookie, CORS, GitHub App, Cloudflare, or session changes include negative-path tests.
• API/OpenAPI/MCP behavior is updated and tested where needed.
• UI changes use live API data or real empty/error/loading states, not production mock/demo fallbacks.
• Visible UI changes include a UI Evidence section below with JPG/JPEG or PNG screenshots arranged as organized, captioned, clickable thumbnails. SVG screenshots are not used as review evidence. Review-only screenshots or recordings are not committed to the repository.
• Public docs/changelogs are updated where needed; changelogs are only edited for release-prep PRs.

The core is a pure string transformer over caller-supplied data — no network, no env, no model call, no publish path. The body-text-only boundary has its own test.

UI Evidence

Not applicable — no UI change.

Notes

• Capture drafting-quality feedback: feed post-merge gaps back into #8103's issue-drafting tool #8118 (drafting-miss feedback loop) extends this core next — the IssueDraftOptions bag is the intended seam.

[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 92.04%. Comparing base (b3cb3c9) to head (39344e0).
⚠️ Report is 2 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #8148   +/-   ##
=======================================
  Coverage   92.03%   92.04%           
=======================================
  Files         764      765    +1     
  Lines       77540    77634   +94     
  Branches    23436    23458   +22     
=======================================
+ Hits        71366    71460   +94     
  Misses       5062     5062           
  Partials     1112     1112           

Flag	Coverage Δ
shard-1	56.72% <100.00%> (-0.92%)	⬇️
shard-2	56.98% <0.00%> (+6.07%)	⬆️
shard-3	48.72% <0.00%> (-5.73%)	⬇️

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

Files with missing lines	Coverage Δ
src/services/issue-drafting.ts	100.00% <100.00%> (ø)

[loopover-orb]

Important

🟪🟪🟪🟪🟪🟪🟪🟪🟪🟪🟪🟪

🔍 LoopOver is reviewing…

AI analysis is in progress. This comment will update when the review is complete.

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