By trevoredris
NASA Power of 10 rules for safety-critical code as agentic skills for Claude Code.
Show the NASA Power of 10 rules index with severities and per-rule skill names.
Review a file, directory, or git diff against all 10 NASA Power of 10 rules. Pass scope as argument (e.g. /pow10-review src/foo.c, /pow10-review main..HEAD).
NASA Power of 10 Rule 1 — Restrict control flow to simple constructs (no goto, setjmp/longjmp, recursion). Severity: blocker.
NASA Power of 10 Rule 2 — Every loop must have a statically determinable upper bound. Severity: blocker.
NASA Power of 10 Rule 3 — No dynamic memory allocation after initialization. Severity: blocker.
NASA Power of 10 Rule 4 — Functions ≤60 lines (one printed page). Severity: high.
NASA Power of 10 Rule 5 — Average ≥2 runtime assertions per function. Severity: high.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
NASA's Power of 10 rules for developing safety-critical code as a Claude Code plugin: 10 per-rule skills, 2 slash commands, and 1 review subagent.
This is not a linter. It teaches the agent the ten rules well enough to apply them during edits, reviews, and authoring across C, Go, Python, Java, and Kotlin. When a hard automated check is wanted, each skill names the right existing tool to invoke (clang-tidy, golangci-lint, ruff, detekt, SpotBugs). We don't reinvent them.
/plugin marketplace add TrevorEdris/power-of-10-rules
/plugin install pow10@pow10
| Surface | What it does |
|---|---|
/pow10-overview | Show the ten rules with severities + waiver convention |
/pow10-review <scope> | Walk all ten rules over a file, dir, or git diff ref |
pow10-rule-NN-<slug> skills | Auto-loaded when relevant code is edited or discussed |
pow10-auditor agent | Same review as /pow10-review, run as a subagent |
.claude-plugin/
├── plugin.json
└── marketplace.json
skills/
├── pow10-rule-01-control-flow/
│ ├── SKILL.md
│ └── references/{c,go,python,java,kotlin}.md
├── pow10-rule-02-bounded-loops/...
├── pow10-rule-03-no-dynamic-memory/...
├── pow10-rule-04-short-functions/...
├── pow10-rule-05-assertion-density/...
├── pow10-rule-06-minimum-scope/...
├── pow10-rule-07-check-return-values/...
├── pow10-rule-08-limited-preprocessor/...
├── pow10-rule-09-restrict-pointers/...
└── pow10-rule-10-warnings-as-errors/...
commands/
├── pow10-overview.md
└── pow10-review.md
agents/
└── pow10-auditor.md
Repo root is the plugin root — no nested plugins/<tool>/ indirection.
When a rule must be broken, leave an inline comment so reviewers and future agents see it:
// pow10: allow rule=N until=YYYY-MM-DD owner=<handle> reason="..."
Use the file's native comment syntax (//, #, --). All four fields required. No statefile.
C, Go, Python, Java, Kotlin. Each rule's per-language section gives concrete violation patterns, remediation snippets, and the existing analyzer to invoke for hard checks.
This repo currently targets Claude Code only. Cursor and Codex variants were removed in 0.2.0 to cut maintenance overhead. If interest emerges, contributors can re-add them under a contrib/ path that mirrors the canonical Claude Code skills.
See CONTRIBUTING.md for the 200-line SKILL.md size policy and the per-language split layout.
MIT. See LICENSE. Rule text paraphrases Holzmann (2006) and the JPL Institutional Coding Standard.
npx claudepluginhub trevoredris/power-of-10-rulesDeterministic Pokemon companion — roll a Pokemon from your user ID with ASCII sprites and shiny rendering
Silent session wiki: automatically captures and persists knowledge from every Claude Code session into a local markdown wiki — zero configuration required.
Spec-driven development workflow: one plugin, seven commands, carrying work from discovery through TDD to review. PRD and roadmap; spec and architecture; slice-by-slice TDD; parallel code/security/a11y review; conventional commits; PR generation.
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Write feature specs, plan roadmaps, and synthesize user research faster. Keep stakeholders updated and stay ahead of the competitive landscape.
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.