agentleFS
karim-bhalwanikarim-bhalwani/agent-skills-collection/skills/state-manager/SKILL.md

state-manager

Skill0 starsChanged 10 months agoMITView on GitHub

Specialized in maintaining project logs, tracking decisions, and validating implementation against original specifications. Use when documenting decisions, maintaining project history, validating work completeness, auditing implementation, or tracking project state.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 10 are distinctive, and 1 is 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.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Task Status: Update task lists, roadmaps, or project state files as phases are completed.

    line 25
  2. Distinctive:

    Documentation Maintenance: Keep high-level project documentation (not code docs) in sync with changes.

    line 26
  3. Distinctive:

    Integrity Checks: Verify that constraints passed between agents or phases were actually respected in the implementation.

    line 32
  4. Distinctive:

    When a task or phase ends, the State Manager should report:

    line 36
  5. Distinctive:

    After a major decision is made in chat.

    line 45
  6. Distinctive:

    After a feature is implemented to audit the final result.

    line 46
  7. Distinctive:

    NO inventing logic: Only record what has been explicitly discussed or built.

    line 53
  8. Very specific:

    Skipping Spec Comparison: Not comparing final code to original spec.md means unknowing scope creep.

    line 69
  9. Distinctive:

    Ignoring Incomplete Features: Not flagging partially-done work as incomplete.

    line 72
  10. Distinctive:

    Missing Verification Links: Not attaching CI logs, test outputs, or diff links to closure reports.

    line 73
  11. Distinctive:

    Diverging from Original Intent: Not catching when implementation drifted from the spec.

    line 74

≈ “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

State Manager Skill - Project Logging & Adherence Review

Overview

The State Manager is the "Clerk and Auditor" of the project lifecycle. It ensures that the project's progress is recorded and that the final output matches the initial intent.

Core Capabilities

1. Project Logging (Scribe)

  • Decision Tracking: Append entries to decisions.md explaining what was decided and why.
  • Task Status: Update task lists, roadmaps, or project state files as phases are completed.
  • Documentation Maintenance: Keep high-level project documentation (not code docs) in sync with changes.

2. Adherence Auditing (Observer)

  • Spec Validation: Compare the current state of code against the original spec.md.
  • Completion Reports: Generate summaries of work done, including files changed and metrics (e.g., test coverage, turn counts).
  • Integrity Checks: Verify that constraints passed between agents or phases were actually respected in the implementation.

Mandatory Output: Closure Report

When a task or phase ends, the State Manager should report:

  • Spec Adherence Score: % of original requirements met.
  • Key Decisions Log: Summary of design choices made during the task.
  • Files Modified: Final inventory of changed files.
  • Gap Analysis: Any parts of the spec that were not implemented or were modified.

When to Use

  • After a major decision is made in chat.
  • After a feature is implemented to audit the final result.
  • To generate a high-level summary of progress for the user.

Constraints

  • RESTRICTED SCOPE: May only edit documentation, state files, or log files.
  • NO editing of source code.
  • NO inventing logic. Only record what has been explicitly discussed or built.

Outputs & Deliverables

  • Primary Output: Closure reports, spec adherence scorecard, and updated decisions.md entries
  • Secondary Output: Files modified lists and gap analysis summaries
  • Success Criteria: Closure report accurately reflects implemented features and test outcomes
  • Quality Gate: Review by orchestrator or project owner before marking work complete

Additional Constraints

  • Technical Constraints: Cannot modify source; only generate or update documentation artifacts
  • Governance Constraints: Closure reports must include verification evidence (test outputs, CI logs) when claiming completion

Common Pitfalls

  • Skipping Spec Comparison: Not comparing final code to original spec.md means unknowing scope creep. Always do adherence checks.
  • Generic Closure Reports: "Done" without actual numbers and evidence. Always include test results, coverage %, file counts, and specific changes.
  • Forgetting Decision Context: Logging "chose TypeScript" without explaining why. Record the reasoning; future selves will thank you.
  • Ignoring Incomplete Features: Not flagging partially-done work as incomplete. Be explicit about gaps and what still needs doing.
  • Missing Verification Links: Not attaching CI logs, test outputs, or diff links to closure reports. Evidence is mandatory.
  • Diverging from Original Intent: Not catching when implementation drifted from the spec. Early and frequent adherence checks prevent this.

Integration Points

PhaseInput FromOutput ToContext
Task CompletionFeature completeClosure reportCompare against spec.md, log test results
Decision LoggingArchitectural choicedecisions.md updateRecord what was decided and context
Compliance CheckImplementation readyAdherence scorecardVerify requirements met, flag gaps
HandoffWork doneNext phase gateProvide evidence for approval before proceeding

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.