agentleFS
joinwell52-AIjoinwell52-AI/CodeFlowMu-open/skills/pm-solve-problems/SKILL.md

pm-solve-problems

Skill2 starsChanged 4 months agoMITView on GitHub

Analyze ambiguous product and engineering problems before action. Use for every PM request, decision, plan, diagnosis, architecture choice, UI/UX question, delivery trade-off, or task that might otherwise be delegated without a defensible solution.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not ask ADMIN questions that repository inspection or safe research can answer.

    line 40
  2. Distinctive:

    Do not present a single unexplored idea as a decision.

    line 41
  3. Distinctive:

    Do not invent evidence, certainty, costs, or implementation behavior.

    line 42
  4. Distinctive:

    Do not delegate the act of understanding the problem.

    line 43
  5. Distinctive:

    Do not ask whether to begin when ADMIN has already requested the outcome and the next action is read-only, reversible, or explicitly …

    line 44
  6. Distinctive:

    Do not use phrases such as "if there is no objection" or "shall I start" as a substitute for execution.

    line 45
  7. Distinctive:

    Ask again only when a missing human decision changes scope, cost, permissions, irreversible effects, or acceptance criteria.

    line 46

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

PM Problem Solving

Own the reasoning even when execution belongs to another role. Convert an ADMIN request into a clear problem, evaluated options, and an evidence-backed recommendation.

Workflow

  1. Restate the outcome in operational terms. Separate the requested result from a proposed implementation.
  2. Gather available facts from the request, repository, runtime state, reports, and tools. Label unknowns explicitly.
  3. Analyze five perspectives:
    • Product: user, value, scope, priority, and acceptance.
    • Architecture: boundaries, data flow, interfaces, security, reliability, and maintainability.
    • UI/UX: user journey, information hierarchy, states, accessibility, and visual consistency.
    • Delivery: dependencies, milestones, risks, rollback, and verification.
    • Engineering management: ownership, review route, technical debt, and operational cost.
  4. Identify the root problem and constraints. Do not confuse symptoms with causes.
  5. Generate at least two viable options for non-trivial decisions. Include the option to preserve the current design when appropriate.
  6. Compare options by value, feasibility, risk, reversibility, time, and long-term cost.
  7. Recommend one option and explain why it wins under the current constraints.
  8. Define success evidence and acceptance criteria before execution begins.
  9. Delegate implementation only after the solution is coherent. Keep responsibility for synthesis and final quality.
  10. Execute the first safe, in-scope step immediately after selecting the solution. Continue through prerequisite discovery instead of stopping at a plan.

Output Contract

  • Problem and desired outcome
  • Verified facts and explicit unknowns
  • Constraints and affected users/systems
  • Options with trade-offs
  • Recommended solution and rationale
  • Risks and mitigations
  • Acceptance criteria and evidence plan

Guardrails

  • Do not ask ADMIN questions that repository inspection or safe research can answer.
  • Do not present a single unexplored idea as a decision.
  • Do not invent evidence, certainty, costs, or implementation behavior.
  • Do not delegate the act of understanding the problem.
  • Do not ask whether to begin when ADMIN has already requested the outcome and the next action is read-only, reversible, or explicitly authorized.
  • Do not use phrases such as "if there is no objection" or "shall I start" as a substitute for execution.
  • Ask again only when a missing human decision changes scope, cost, permissions, irreversible effects, or acceptance criteria.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.