agentleFS
Sign inStart free
bordenetbordenet/superpowers-plus/skills/observability/completeness-check/skill.md

completeness-check

Skill8 starsChanged 7 months agoMITView on GitHub

Detect incomplete work in repositories from AI assistant crashes, context exhaustion, or mid-implementation distractions. Use before claiming work complete or when auditing accumulated debt.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    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:

    MANDATORY: Use this skill before claiming work is complete on ANY repo, or when auditing accumulated incomplete work.

    line 27
  2. Distinctive:

    Before marking a task as complete

    line 32
  3. Distinctive:

    After implementing a multi-file change

    line 33
  4. Distinctive:

    When a code review requests "check all callers"

    line 34
  5. Distinctive:

    All Critical resolved · score ≥70 · no regressions · deferred findings documented.

    line 149

≈ “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

completeness-check

Wrong skill? Pre-commit checks → pre-commit-gate. Output inspection → output-verification. Repo health → holistic-repo-verification.

MANDATORY: Use this skill before claiming work is complete on ANY repo, or when auditing accumulated incomplete work.

When to Use

  • Quick sanity check that nothing was missed in a changeset
  • Before marking a task as complete
  • After implementing a multi-file change
  • When a code review requests "check all callers"

The Core Principle

AI coding assistants frequently leave incomplete work when they:

  • Get distracted mid-implementation
  • Crash or lose context
  • Context-switch away from unfinished tasks
  • Exhaust their context window (~70% utilization shows degradation)

This leaves humans out of the loop. This skill helps detect and surface that incomplete work.

Detection Categories

CategorySeverityConfidenceDetection Method
Broken internal links🔴 CriticalHighRegex + file existence
Retired file references🔴 CriticalHighGrep + file existence
Hallucinated dependencies🔴 CriticalHighRegistry verification
Security-critical TODOs🔴 CriticalHighGrep + keyword filter
Generic TODO/FIXME🟡 WarningHighGrep
Debug statements🟡 WarningHighGrep
Commented-out code🟡 WarningLowHeuristic
Missing tests (coverage)🟡 WarningHighCoverage tools
Missing tests (pairing)🟡 WarningMediumHeuristic
Placeholder content🟡 WarningLowRegex
Stale documentation🟡 WarningMediumDiff vs reality
Context exhaustion markers🟡 WarningMediumPositional analysis
Happy path bias🟡 WarningMediumError handling density
Deployment gaps🟡 WarningMediumMissing config/Docker
Missing docstrings🟢 InfoMediumAST/Regex
Empty directories🟢 InfoHighFind
Orphaned files🟢 InfoMediumImport graph
Dead code🟢 InfoHighLinter output

CLI Usage

# Proactive mode (before claiming work complete)
completeness-check --mode proactive

# Reactive mode (auditing accumulated incomplete work)
completeness-check --mode reactive --verbose

# Target specific directory
completeness-check ./src --mode proactive

# Output formats
completeness-check --format summary    # Score + counts only
completeness-check --format detailed   # Full findings with file locations
completeness-check --format json       # Machine-readable for CI integration

Scoring Formula

Base deductions:
  Critical (High conf):   15 points each
  Critical (Med/Low):     10 points each
  Warning (High conf):     7 points each
  Warning (Med/Low):       5 points each
  Info:                    1 point each

Score = max(0, 100 - deductions)
ScoreInterpretation
90-100✅ Clean: minimal incomplete work detected
70-89🟡 Light: some loose ends, minor cleanup needed
50-69🟠 Moderate: noticeable gaps, review recommended
30-49🔴 Heavy: significant incomplete work, prioritize cleanup
0-29⛔ Severe: repo appears abandoned mid-implementation

Output Format: Summary

╭─────────────────────────────────────────────────────────────╮
│  COMPLETENESS CHECK                                         │
│  Repository: genesis                                        │
│  Mode: Proactive                                            │
╰─────────────────────────────────────────────────────────────╯

Completeness Score: 73/100 (Light)

Summary:
  🔴 Critical: 2
  🟡 Warning:  12
  🟢 Info:     5

Implementation Phases

Phase 1: Fast (grep-based)

  • TODO/FIXME comments
  • Debug statements
  • Placeholder patterns
  • Empty directories

Phase 2: Tool-Based

  • ESLint/staticcheck for dead code
  • Coverage tools for missing tests
  • Package registry for hallucinated deps

Phase 3: Deep (expensive)

  • AST parsing for orphaned code
  • Semantic analysis for stale docs
  • Cross-file dependency graphs

Success Criteria

All Critical resolved · score ≥70 · no regressions · deferred findings documented.

Failure Modes

FailureRecovery
Self-grading bias — claiming 100% without independent verificationCross-check against original scope, not your own checklist
Scope creep — adding items to inflate completeness percentageCompare checklist against original request verbatim
Surface-level scan — checking filenames onlyRead file contents. Check for TODO, FIXME, incomplete implementations
Missing abandoned branchesCheck git branch -a for stale feature branches
False positive on intentional stubsCheck git history — recent stubs may be in-progress, not abandoned
Skipping error handling/cleanup checksCheck error paths, rollback logic, and cleanup — not just happy path

Companion Skills

  • exhaustive-audit-validation: Deeper audit (heavier than this skill)
  • verification-before-completion: Pre-completion gate
  • holistic-repo-verification: Full repo health check
  • measurement-integrity: Metric validation

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.

Check your setup, free