align
Skill0 starsChanged 2 months agoApache-2.0View on GitHub
Compute the cross-corpus ALIGNMENT scorecard — a deterministic, no-LLM roll-up that scores how well the ontology bridges CODE ↔ DOC ↔ ADR ↔ TEST bi-directionally, surfaces the Test-Traceability Ledger (no orphan tests; every test gates an ontology relationship), and emits a Contract-Test Readiness verdict + ranked actionable items + a trend series. Run it as a periodic heartbeat to see what blocks moving forward to contract tests. Writes lineage/{repo}/alignment-scorecard.{md,yaml}.
17 other projects use a version of this.
Adopted by
Most recently taken up by
- FooJiaYin/dev-skillstheir own version, first seen Nov 16, 2026
- dimsedra/skillstheir own version, first seen Nov 8, 2026
- devjiwonchoi/skillstheir own version, first seen Nov 8, 2026
- KerberosClaw/kc_ai_skillstheir own version, first seen Nov 1, 2026
- lgy1027/specflow-agent-packtheir own version, first seen Nov 1, 2026
- hmmhmmhm/aligns.mdtheir own version, first seen Oct 17, 2026
Used with
Projects with it also use blueprint and research.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
WordPressSkillblueprint
Use when the deliverable is WordPress Playground Blueprint JSON or a Blueprint bundle, including creating, editing, reviewing, validating schema keys, choosing steps/resources, and debugging Blueprint files. For only running or sharing a Playground environment, use wp-playground.
Also in 3 projects that use this
NVIDIASkillresearch
Before designing a feature, research how comparable libraries solve the problem, check best practices, known pitfalls, and upstream changes. Use when starting a new feature, investigating an unfamiliar area, or the user asks to research something.
Also in 3 projects that use this
Other versions17 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 17
- ambient-code/platform★ 131same name
- KerberosClaw/kc_ai_skills★ 79same name
- samirpatil2000/skills★ 17same name
- hmmhmmhm/align.md★ 11same name
- hmmhmmhm/aligns.md★ 11same name
- HaipingShi/coderail★ 8same name
- JustineDevs/meta-architect★ 5same name
- lgy1027/specflow-agent-pack★ 4same name
- devjiwonchoi/skills★ 2same name
- dimsedra/skills★ 2same name
7 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository27 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- code-walkSkill
- conceptsSkill
- contributeSkill
- coverageSkill
- doc-gap-checkSkill
- enrichSkill
- enumerateSkill
- find-implicit-adrsSkill
- implementSkill
- ingest-docsSkill
- log-issueSkill
- navigateSkill
- next-batchSkill
- orientSkill
- panelSkill
- probe-defineSkill
- probe-runSkill
- probeSkill
- reflect-featureSkill
- retrieveSkill
- reviewSkill
- scanSkill
- statusSkill
- streamsSkill
- test-coverageSkill
- triageSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 16 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 34
The honesty rule (hard-coded): every alignment metric is reported as aligned / checked / total.
- Very specific: line 61
Read back lineage/{repo}/alignment-scorecard.md and relay to the maintainer: the readiness verdict, the blockers, the per-dimension grades, …
- Very specific: line 64
If any dimension is RED for a *new* reason vs the prior trend row, call it out as a regression.
“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Alignment scorecard (/align)
The cross-corpus consistency dashboard. A deterministic roll-up — no LLM, no subagents in default mode — over the already-derived ontology + ground-truth anchors. It answers the maintainer's standing question:
What is the level of alignment between documentation, ADRs, the codebase, and the tests — and what blocks us from building the contract tests that pin doc-claims to implementation?
It does not re-derive anything (that's what /enrich, /concepts,
/find-implicit-adrs, /doc-gap-check, /test-coverage, /reflect-feature
produce). It joins those outputs + the graph, computes bi-directional
coverage ratios, and scores them. The expensive agentic re-verification is the
separate --deep mode (phase 2).
What it measures (5 dimensions)
| Dim | Question |
| A Ontology ↔ Code | is the ontology a fresh, faithful mirror of code? (substrate freshness, enrichment, axis integrity) |
| B Ontology ↔ Doc (bi-dir) | if docs claim it, does code exist? (DESCRIBES→code resolution) · if implemented, is it documented? (features with inbound DESCRIBES) · open DocGaps · doc drift |
| C Ontology ↔ ADR | published ADRs ingested · REALISES code links · candidate disposition |
| D Test-Traceability Ledger | no orphan tests (every test/gap carries a typed gate) · ADRs enforced · features validated · bugs/scopes regress-guarded · probe execution + named-integration stacks |
| E Trust meta-gate | graph freshness · embeddings on? · latest /panel verdict · reflection coverage (the discount on every alignment claim) |
The honesty rule (hard-coded): every alignment metric is reported as
aligned / checked / total. A low contradiction count is rendered as
"UNKNOWN over N%", never "aligned" — because the layer that proves alignment
(feature-reflection) has run on <1% of features. The score cannot hide a blind
spot. The top line is a Contract-Test Readiness verdict (NOT-READY /
PILOT-READY / READY) + explicit blockers + the ready-now subset (what is
fully bridged and contract-testable today).
The Test-Traceability Ledger implements feedback_tests_as_deterministic_gates:
a test is never a free-floating item — it enforces→ADR / validates→Feature /
regresses→Finding-Issue / guards→RefactoringScope, checked bidirectionally
(every test has a gate; every test-worthy subject has a test or a gated gap).
Argument forms
| Form | Behaviour |
/align [<repo>] | Default (repo defaults to odd-platform). Compute the scorecard from the canonical files NOW (fresh load_substrate + project, independent of the last graph-build), write alignment-scorecard.{md,yaml} (the .yaml appends one trend row), print the summary + the top action. Cheap + idempotent — the periodic heartbeat. |
/align --show [<repo>] | Read-only. Print the existing lineage/{repo}/alignment-scorecard.md (no recompute). |
/align --json [<repo>] | Emit the machine payload (the written YAML) to stdout for downstream tooling. |
/align --deep [<repo>] | Phase 2 — not yet built. The agentic contract audit: fan out subagents to sample-verify the cheap signals — re-fire reflect-feature on the highest-value un-reflected features, sample Doc→Code claims for live verification, drift-check each ADR's REALISES. Each finding logged as a gated TestGap via playbooks/follow-up-on-disk.md. Until built, the skill says so and runs the default mode. |
Procedure (default mode)
- Run
uv run --project lineage/_extractor lineage-extractor alignment <repo>(deterministic; ~15 s; reads the canonical files + the built graph'sbuild-info.yaml+ git HEAD + the published ADR pages + the probe corpus). - Read back
lineage/{repo}/alignment-scorecard.mdand relay to the maintainer: the readiness verdict, the blockers, the per-dimension grades, the ready-now subset, and the top 3 actionable items. - If any dimension is RED for a new reason vs the prior
trendrow, call it out as a regression. The trend series is the periodic-monitoring surface.
When to run
- Per sprint checkpoint / milestone — alongside
/panel(which audits the methodology;/alignaudits the ontology's fidelity — complementary, and/alignconsumes the latest/panelverdict in its trust gate). - After a batch of
/enrich+ reducers — to see whether coverage moved the alignment numbers. - Before starting contract-test work — the readiness verdict + ready-now subset tell you exactly where to aim, and the actionable items name the blockers (Phase-4 Test layer, reflection coverage, ADR re-ingest, etc.).
On-demand only — no daemon, no schedule (APPROACH.md Rule 12; local-only).
Outputs
lineage/{repo}/alignment-scorecard.md— human, diffable, committed.lineage/{repo}/alignment-scorecard.yaml— machine metrics +trend:series (last 50 runs), committed.
Cross-references
lineage/_extractor/src/lineage_extractor/alignment.py— the implementation.feedback_tests_as_deterministic_gates(memory) — the Ledger's spine.adrs/drafts/ground-truth-lineage.md— Phase 4 (theTestlayer the readiness verdict gates on).lineage/GRAPH-TOPOLOGY.md— the labels + edges rolled up here./panel— the complementary methodology meta-review.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

