Add a complementary soft layer to the hard gates: advisory_flags surface
governance hints (e.g. oversized-but-not-blocked PRs -> suggest splitting) in
the scorecard WITHOUT affecting score or verdict. Hard gates stay the real
teeth; advisory leaves the call to humans.
- additive & backward-compatible: default off; policies lacking the section
behave exactly as before (verified: default policy on real PR #275 yields
zero advisory section, PASS 87/100 unchanged)
- deterministic: consumes only already-collected PR data (no wall-clock input),
preserving same-input -> same-score -> same-verdict
- first flag large_pr_files (changed-files >= threshold) answers the
166-open-PR backlog reality where big PRs should be flagged, not blocked
- evaluate_advisory_flags() is separate from score_dimensions/decide_verdict;
the 5-dim scoring and verdict logic are untouched
- 7 new unit tests (15 total) incl. an invariant that enabling advisory does
NOT change total/verdict; verified live on real open PR #275
- REFERENCE.md 1.9 + gatekeeper.yaml + verification.md E + demo scorecard
Extends the merged gatekeeper line (#90 skill -> #219 workflow -> #220 issueops).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A Go workflow (subtopic 3) that composes existing gitlink-cli commands
into a reproducible research-team insight pipeline:
[search +users] -> profile statistics x N (+ user +info)
-> aggregate (ability matrix / discipline coverage / roles / languages)
-> render team report (self-contained HTML, inline SVG: ability radar
overlay + collaboration network graph)
-> archive (team-insight.md + command_log.json)
Implemented in Go to match the main repo's stack (shared go.mod, covered
by go build / vet / test / gofmt); standard library only, no third-party
deps. Handles Windows .cmd CLI shims, retries transient TLS timeouts, and
auto-flags mirror-inflated accounts in the report.
Includes unit tests, docs (architecture + runbook + SVG diagram),
run_demo.ps1, sample config, and real demo outputs (3 GitLink members,
19 command invocations). Reuses the gitlink-research-profile skill's
portrait model.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Runnable reference implementation of the merged gitlink-gatekeeper Skill:
collect -> route -> decide (deterministic 0-100 scorecard, three-state verdict)
-> write-back (comment + label + tracking issue, --apply gated, never auto-merge).
- scripts/gatekeeper_workflow.py: single-PR gate loop (pure stdlib, py>=3.9)
- scripts/gatekeeper_sweep.py: repo-wide triage — real run over all 113 open PRs
of this repo (read-only): PASS 105 / COMMENT 6 / REQUEST_CHANGES 2, avg 88.5
- findings.example.json: real code-review findings of a real PR diff (verifiable line numbers)
- ci-example/: Gitea Actions integration (exit code 2 = REQUEST_CHANGES as CI gate)
- docs: architecture / quickstart / runbook / verification (real-platform evidence)
- tests/test_scoring.py: determinism regression guard (green)
Closes the sub-task-3 loop on top of the label command (#89) and the
gitlink-gatekeeper Skill (#90), sharing one policy file across all three layers.
wbtiger
maidamaliziasimnw
Angel123456
zhangqing
whzy
Ct201314
Jiachen Li