Agent skill

documentation-standards

Apply professional documentation standards. Use when writing README files, commit messages, code comments, technical docs, or any user-facing text. Enforces evidence-based claims, no marketing language.

Stars 0
Forks 0

Install this agent skill to your Project

npx add-skill https://github.com/mojoatomic/triton/tree/main/.claude/skills/documentation-standards

SKILL.md

Documentation Standards

Technical documentation must be factually accurate, conservative in claims, and free of marketing language.

Formatting Rules

No emojis. No checkmarks, party poppers, or unicode decorations. Use plain text: "Compliant", "Implemented", "Supported", "Pending".

No excessive formatting. Use bold sparingly for genuine emphasis only.

Prohibited Language

Marketing Hyperbole - Never Use

  • "Production-ready"
  • "Enterprise-grade"
  • "Best-in-class"
  • "Industry-leading"
  • "State-of-the-art"
  • "Cutting-edge"
  • "Revolutionary"
  • "World-class"
  • "Unparalleled"
  • "Seamless"
  • "Robust" (without specific meaning)
  • "Powerful" (without quantification)

Absolute Claims - Never Use

  • "Perfect"
  • "Flawless"
  • "100% accurate"
  • "Guaranteed"
  • "Always works"
  • "Never fails"
  • "Completely secure"

Competitive Claims - Never Use

  • "Better than X"
  • "Superior to alternatives"
  • "More accurate than other tools"
  • "Faster than existing solutions"

Vague Endorsement - Never Use

  • "NASA-approved" (unless actually approved)
  • "Certified accurate"
  • "Validated by experts"
  • "Industry-standard"

Required Patterns

Evidence-Backed Statements

Bad Good
"Highly accurate" "Median error: 1.4 arcseconds across 7 bodies"
"Fast performance" "Typical response time: 25-75ms"
"Well-tested" "33 unit tests, 6 test suites, 100% pass rate"
"Production-ready" "Deployed in [context]; further testing recommended for [other contexts]"

Bounded Statements

Bad Good
"Low latency" "< 100ms typical, < 500ms worst case"
"High precision" "1.4 arcsec median, 8.6 arcsec maximum"
"Comprehensive coverage" "87% line coverage, 92% branch coverage"

Honest Limitations

Always include a limitations section stating what the software does NOT do.

Specific Methodology

When describing validation: state comparison target, methodology, dataset, and link to reproducible tests.

Commit Messages

No emojis. Use conventional commits format:

feat: add depth controller PID implementation
fix: resolve sleep_us variable shadowing
docs: update validation methodology section
test: add unit tests for state machine transitions

Do not editorialize. No "awesome", no exclamation points.

Code Comments

Comments explain WHY, not WHAT. No cheerleading.

c
// Bad
// This awesome function does the magic!

// Good  
// Apply temperature compensation per MS5837 datasheet section 4.2.

No TODO without ownership:

c
// Bad
// TODO: fix this later

// Good
// TODO(doug): Add timeout handling - see issue #47

README Structure

markdown
# Project Name

One-sentence factual description.

**Status:** Development / Testing / Stable
**Validation:** [Specific methodology and results]

## What This Does

[Factual description]

## Limitations

[Explicit statement of what it doesn't do]

## Validation

[Methodology, data sources, reproducible tests]

## Known Issues

[Current problems]

Pull Request Descriptions

State what changed and why. No self-congratulation.

markdown
## Add Core 1 health monitoring

Adds heartbeat-based detection of Core 1 stalls.

Changes:
- Add g_core1_heartbeat volatile counter
- Add check_core1_health() to safety monitor
- Add EVT_CORE1_STALL event code

Testing:
- Unit test: mock heartbeat stall, verify emergency trigger

Summary

Write as if reviewed by a skeptical engineer who will challenge every unsubstantiated claim.

Expand your agent's capabilities with these related and highly-rated skills.

mojoatomic/triton

power-of-10

NASA/JPL Power of 10 coding rules for safety-critical C code. Use when writing embedded C firmware, safety-critical systems, aerospace code, or any project requiring high reliability. Enforces static analysis, bounded loops, no dynamic memory.

0 0
Explore
mojoatomic/triton

github-hygiene

Enforces strict GitHub workflow hygiene and best practices. Use this skill for ALL git and GitHub operations including commits, branches, pull requests, and repository management. Core rules - NEVER commit directly to main/master, ALWAYS work on feature branches, use GitHub MCP for all GitHub API operations (issues, PRs, reviews), require PRs for all merges. Triggers on any git command, GitHub operation, or when starting new feature work.

0 0
Explore
mojoatomic/triton

safety-critical-embedded

Safety-critical embedded system design patterns. Use when building dual-core safety monitors, failsafe systems, watchdog implementations, or any system where failure could cause harm. Covers isolation, fault detection, and emergency procedures.

0 0
Explore
mattpocock/skills

handoff

Compact the current conversation into a handoff document for another agent to pick up.

111,310 9,758
Explore
mattpocock/skills

setup-pre-commit

Set up Husky pre-commit hooks with lint-staged (Prettier), type checking, and tests in the current repo. Use when user wants to add pre-commit hooks, set up Husky, configure lint-staged, or add commit-time formatting/typechecking/testing.

111,310 9,758
Explore
mattpocock/skills

obsidian-vault

Search, create, and manage notes in the Obsidian vault with wikilinks and index notes. Use when user wants to find, create, or organize notes in Obsidian.

111,310 9,758
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results