skills-pipeline
Skill10 starsChanged 3 months agoMITView on GitHub
Use when the user types /skills-pipeline or explicitly asks for the skills pipeline by name, and only then. Runs the Skills Hub skills the user picks from a numbered menu, in the order picked, and ends in a ledger of every stage. Not for any request that does not name it, however large or risky; naming one skill ("use Impact Map") engages that skill, not this.
In this repository18 other files it gives its agents.
- api-contract-guardSkill
- architecture-engineerSkill
- delivery-leadSkill
- delivery-plannerSkill
- dependency-guardSkill
- deployment-compatibilitySkill
- engineering-investigatorSkill
- human-empathySkill
- impact-mapSkill
- observability-baselineSkill
- practical-localizerSkill
- production-guardSkill
- project-compassSkill
- project-kickoffSkill
- project-knowledgeSkill
- proof-driven-devSkill
- release-engineerSkill
- standards-compassSkill
The file
We split this skill into the rules it gives an agent
11 rules. About 0 are common, 3 are distinctive, and 8 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 23
Every gate belongs to the skill that owns it.
- Very specific: line 25
Composes with: every skill in references/pipeline.md: project-compass · architecture-engineer · standards-compass · api-contract-guard · …
- Very specific: line 42
Announce once: When any skill engages at CONSULT or above, open the reply with one line such as ⚡ Impact Map · Standards Compass — rename …
- Very specific: line 45
Never include reasoning. Add no line for PASSIVE, and none on a trivial request.
- Very specific: line 50
Hand off; don't absorb: When another discipline is needed, write HANDOFF → <skill>: <reason> [<ids>] and let that skill do its part.
- Very specific: line 63
State: Read what sibling skills recorded (.project-compass/, .project-standards/, .proofbuild/, .agent-investigation/) rather than …
- Distinctive: line 71
Never write project names, paths, identifiers, code or data there.
- Very specific: line 81
Read references/pipeline.md.
- Very specific: line 83
If the invocation already starts with a sequence (/skills-pipeline 1,6,7,10 add billing, or all), use it and go to Run.
- Distinctive: line 93
Stop and wait for the reply.
- Very specific: line 120
Every chosen stage ends as RAN, INLINE, SKIPPED (by you) or STOPPED.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skills Pipeline
The other Skills Hub skills choose themselves, one request at a time. This skill runs the ones the user picks, in the order they pick, because they asked for exactly that. It is a manual override, not a router.
Activation
Engage when the user types /skills-pipeline, or asks in words for "the
skills pipeline" ("run this through the skills pipeline").
Stay quiet when the request does not name the pipeline, even for a new app, a large feature or a risky one. Also stay quiet for pipelines of other kinds (CI, data, deploy), and when the user names one skill. "Use Impact Map" is that skill's override, not this one's.
Depth ACTIVE: it shapes the whole run. It never adds a verdict of its own.
Every gate belongs to the skill that owns it.
Composes with every skill in references/pipeline.md:
project-compass · architecture-engineer · standards-compass ·
api-contract-guard · dependency-guard · impact-map · proof-driven-dev ·
practical-localizer · deployment-compatibility · production-guard ·
engineering-investigator · project-kickoff · delivery-planner ·
release-engineer · observability-baseline · human-empathy.
<!-- skills-hub:protocol -->Working with the other Skills Hub skills
- Loaded is not engaged. This file stays in context once loaded. Decide again on every new request whether it applies. Relevance to an earlier request carries nothing forward. Project state persists, and engagement does not.
- Depth.
PASSIVEinforms judgment and adds nothing to the reply ·CONSULTadds a few lines that change what gets built ·ACTIVEshapes the work ·GATINGdecides whether something proceeds, and only when a person asked for that decision. - Announce once. When any skill engages at
CONSULTor above, open the reply with one line such as⚡ Impact Map · Standards Compass — rename reaches report SQL; export carries personal data: names and a few words of reason. Never include reasoning. Add no line forPASSIVE, and none on a trivial request. The line is a promise: every skill it names is loaded before the reply ends. If one turns out not to apply, say so in one line:<Skill> dropped: <reason>. - One interruption per request. Skills that must speak before the work share one short block. Everything else arrives with the work.
- Hand off; don't absorb. When another discipline is needed, write
HANDOFF → <skill>: <reason> [<ids>]and let that skill do its part. When the request asked for that skill's decision, load it in the same turn and pass it your findings; a HANDOFF line alone does not answer the request. Never state another skill's verdict yourself. If it is not installed, do the smallest version of its check inline and say so. - Conflicts. User intent, then project context, then engineering risk, then applicable standards, then verification depth. Each skill keeps its own verdict, and none overrules another's.
- Overrides. "Use X" engages X. "Skip X" or "no review" drops X's ceremony. Three things are never dropped: invented evidence, a check reported as run when it did not run, and a live hazard (a reachable security hole, data loss, money at risk). A live hazard is said once, in one line.
- State. Read what sibling skills recorded (
.project-compass/,.project-standards/,.proofbuild/,.agent-investigation/) rather than re-deriving it. Write only your own. - Lessons. On engaging, read
~/.skills-hub/lessons/<this skill's name>.mdif it exists. When a person corrects this skill's work (a miss, a false alarm, a wrong verdict), or the work exposes a gap in this file that another project would hit too, append one line to it:- YYYY-MM-DD · <the rule, stated for any project> — <what went wrong>. Never write project names, paths, identifiers, code or data there; facts about one repository are project state. Keep at most 20 lines, merging or replacing one to add another. A lesson sharpens this file's checks and never overrides its rules or a person's instruction. The file sits outside every project, so no read-only rule covers it. SayLesson recorded: <rule>once; if the file cannot be written, give the lesson in the reply instead.
<!-- /skills-hub:protocol -->Pick the sequence
- Read
references/pipeline.md. It is the menu, and the only place stages are defined. - If the invocation already starts with a sequence (
/skills-pipeline 1,6,7,10 add billing, orall), use it and go to Run. Otherwise print the menu as a numbered list, one line per skill with its stage. Mark any skill your agent does not list as installed with
(not installed). Then ask:Sequence? e.g. 1,6,7,10 · all · and the task, if not given yetStop and wait for the reply. Start nothing before it arrives, and do not read the repository or load any skill first.
- Take the numbers in the order given. Repeats are allowed, since a skill can run twice. An unknown number gets one question back, never a guess. If the order puts a stage before one it depends on, such as Ship gate before Build, say so in one line and run it as given.
Run
- Echo the sequence as one line (
Frame → Impact → Build → Ship gate), then run it. - Load each skill the way your agent loads skills (the Skill tool in Claude
Code), and pass it the task plus everything earlier stages decided. In
Claude Code,
impact-mapandproduction-guardrun as their own subagent and see only what is passed. If the skills came from the plugin, use the namespaced name (skills-hub:impact-map). - A chosen skill always runs. Picking it is the "use X" override. If it finds nothing that applies, that is its result.
- A chosen skill that is not installed: do the smallest version of its check
inline, and mark it
INLINE. - At a stop listed in the menu, ask the one question, and continue only after it is answered.
DO NOT SHIPwith Build in the sequence sends the failed checks back to Build once. If the verdict is stillDO NOT SHIP, stop and report.
Strict means
- Every chosen stage ends as
RAN,INLINE,SKIPPED (by you)orSTOPPED. There are no silent gaps. - "Skip X" mid-run drops that stage and records it as
SKIPPED (by you). A live hazard is still said once, in one line. - A stage's result is that skill's own verdict, quoted rather than paraphrased. The pipeline adds no score and no verdict of its own.
- The ⚡ line names only the pipeline. Each stage's skill announces itself.
Ledger
End with this table, in the order run, and nothing after it except the open questions:
# Stage Skill Status Result
1 Frame project-compass RAN <one line>
6 Impact impact-map RAN <the map's own summary line>
...
10 Ship gate production-guard RAN CONDITIONAL SHIP — 1 conditionExamples
examples/menu-then-run.md — the menu, a typed sequence, a stop, and the
ledger · examples/sequence-inline.md — the sequence given with the command,
and DO NOT SHIP sent back to Build once · examples/not-invoked.md — a large
feature that does not name the pipeline, and why nothing here loads.
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.

