pm-evolve-skills
Skill2 starsChanged 4 months agoMITView on GitHub
Discover, learn, validate, and retain capabilities that the PM needs but does not yet have. Use whenever a request needs unfamiliar domain knowledge, a missing tool or workflow, a new MCP capability, a new Skill, or stronger evidence than the PM can currently produce.
In this repository42 other files it gives its agents.
- architect-boundary-reviewSkill
- architect-decision-recordSkill
- architect-system-designSkill
- browser-playwright-checkSkill
- browser-useSkill
- code-search-navigationSkill
- controlled-emergence-observerSkill
- dev-code-locationSkill
- dev-small-scope-patchSkill
- dev-test-and-explainSkill
- eval-observation-writingSkill
- eval-promotion-adviceSkill
- eval-risk-gap-analysisSkill
- ops-log-diagnosisSkill
- ops-runtime-healthSkill
- ops-stuck-workflowSkill
- pm-acceptance-criteriaSkill
- pm-architecture-reviewSkill
- pm-delivery-planSkill
- pm-long-horizon-planningSkill
- pm-priority-triageSkill
- pm-product-design-briefSkill
- pm-product-requirementsSkill
- pm-scope-controlSkill
- pm-solve-problemsSkill
- pm-tech-scopeSkill
- qa-regression-checkSkill
- qa-reproduce-issueSkill
- qa-verify-fixSkill
- run-local-command-checkSkill
- test-selectionSkill
- tm-delivery-governanceSkill
- tm-review-routingSkill
- tm-risk-dependencySkill
- ui-information-architectureSkill
- ui-requirementsSkill
- ui-usability-acceptanceSkill
- ui-visual-consistencySkill
- web-extractSkill
- web-researchSkill
- web-searchSkill
- windows-useSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 10 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈10
Distinctive · estimate
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
- Distinctive: line 8
Treat capability gaps as work to resolve, not as a reason to guess or immediately hand the problem back to ADMIN.
- Very specific: line 18
Use skill_search to search the local library, GitHub, and the public web when the capability is not already known.
- Very specific: line 21
Use skill_learn with selected local skill ids and source URLs to build a traceable learning pack.
- Distinctive: line 23
Verify the result against an explicit expectation.
- Distinctive: line 26
Update or create a focused Skill when the procedure will recur.
- Distinctive: line 27
Add a template or example when structure matters.
- Very specific: line 29
Use skill_publish only after a real validation succeeds.
- Distinctive: line 37
Check the relevant UI playbook packages:
- Distinctive: line 44
If the pattern will recur, refine the Playbook or create a smaller reusable template after validation.
- Distinctive: line 66
Do not equate reading with learning.
- Distinctive: line 67
Do not install or create broad capabilities when a smaller local procedure suffices.
- Distinctive: line 68
Do not claim a tool exists until its registry or schema is inspected.
- Distinctive: line 69
Do not fabricate successful experiments or conceal failed attempts.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
PM Skill Evolution
Treat capability gaps as work to resolve, not as a reason to guess or immediately hand the problem back to ADMIN.
Workflow
- State the required capability precisely. Distinguish missing knowledge, missing procedure, missing tool access, and missing permission.
- Inventory what is already available:
- MCP tools and their schemas.
.codeflowmu/agent-skills.manifest.json.- Relevant
skills/*/SKILL.mdpackages. - Repository code, tests, docs, examples, and prior evidence.
- Use
skill_searchto search the local library, GitHub, and the public web when the capability is not already known.
- Select the smallest credible learning source. Prefer source code, tests, official documentation, and proven local examples.
- Load only the relevant Skill or documentation. Do not flood context with the whole catalog.
- Use
skill_learnwith selected local skill ids and source URLs to build a traceable learning pack.
- Use
- Learn by doing: run a minimal experiment, build a small example, inspect a real artifact, or execute a reversible probe.
- Verify the result against an explicit expectation. Record failures and uncertainty honestly.
- Apply the capability to the current problem.
- Retain reusable learning:
- Update or create a focused Skill when the procedure will recur.
- Add a template or example when structure matters.
- Record a proposal when installation, permission, or architectural change requires approval.
- Use
skill_publishonly after a real validation succeeds. Include validation evidence and source URLs; do not overwrite an existing Skill implicitly.
- Reuse retained knowledge on later tasks instead of rediscovering it.
Product / UI capability gaps
Weak product planning is a capability gap. If PM cannot confidently define product positioning, user value, information architecture, interaction flow, visual direction, mobile/PWA/Gateway delivery boundary, or acceptance evidence, PM must:
- Load
pm-product-design-brief. - Check the relevant UI playbook packages:
ui-requirementsui-information-architectureui-visual-consistencyui-usability-acceptance
- Inspect nearby successful product examples or reports before dispatch.
- Produce a concrete design brief, then dispatch DEV / QA / OPS.
- If the pattern will recur, refine the Playbook or create a smaller reusable template after validation.
Capability Decision
available: load and use the existing capability.learnable_now: study and validate it within the current task.creatable: create a focused Skill, helper, or template and validate it.installable: use the approved plugin/skill installation path.permission_blocked: explain the exact permission or ADMIN decision required.not_yet_verified: do not claim competence; continue research or report the limitation.
Evidence Contract
- Capability required
- Sources inspected
- Skill or tool selected
- Validation performed
- Result and remaining uncertainty
- Reusable artifact retained, when applicable
Guardrails
- Do not equate reading with learning; validation is required.
- Do not install or create broad capabilities when a smaller local procedure suffices.
- Do not claim a tool exists until its registry or schema is inspected.
- Do not fabricate successful experiments or conceal failed attempts.
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.

