Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/a7um/zero-review/auto-devnpx skills add A7um/zero-review --skill auto-devgit clone --depth 1 https://github.com/A7um/zero-reviewWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00039 | $0.02409 |
| Opus 5 | $0.00019 | $0.01205 |
| Sonnet 5 | $0.00008 | $0.00482 |
| Haiku 4.5 | $0.00004 | $0.00241 |
Grade A, and why
auto-dev scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Auto-Dev Skill
WHEN TO USE: Task requires building, modifying, or fixing software, tools, or systems and requirements are clear (confidence: MEDIUM or HIGH).
Development Philosophy
Think like an engineer, not a code generator. Quality comes from understanding, not speed.
① Define done — Before touching code, articulate what success looks like. What should work? What must not break? What is out of scope? This is your anchor for every decision that follows. Without it, scope drifts silently.
② Understand before you change — Read the code you're about to modify. Read its callers, its callees, its tests. Map the blast radius. Your change quality is bounded by your understanding quality — when things go wrong, it's almost always because you understood too little, not because you coded too little.
③ Present, then act — After understanding, describe what you found and what you plan to do before doing it. The most productive sessions are those where design is discussed before code is written. The worst sessions are those where the agent reads files and immediately starts editing.
④ One correct change — Re-read the file before each edit. Make one comprehensive change, not ten incremental patches. If you find yourself editing the same file a third time, stop — you don't understand it well enough yet.
⑤ Every result is evidence — A test failure is not "retry with a tweak." Read the assertion. What does it actually say? A passing test suite is not "done" — does it exercise the behavior you changed? An error message is not "try something else" — what specifically went wrong?
⑥ Check against ①, stop — When the success criteria from ① are met, deliver. Don't improve what wasn't asked for. Don't refactor adjacent code. Don't absorb new scope mid-task. If a new request arrives, explicitly decide: defer, replace, or acknowledge the expansion.
Task Classification
| Signal | Paradigm |
|---|---|
| New project/system, no existing codebase | dev/architecture-first |
| Defect, regression, incorrect behavior | bugfix/hypothesis-driven |
| Single function/component, fits existing architecture unchanged | addition/lightweight |
| Everything else (feature addition, behavior extension) | enhancement/delta-design |
What ships with it
34 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- .gitignore 49 B
- config/defaults.json 940 B
- CONTRIBUTING.md 1.7 KB
- examples/code-design-output.md 5.7 KB
- examples/code-review-output.md 1.2 KB
- paradigms/_template.md 446 B
- paradigms/addition/lightweight.md 2.2 KB
- paradigms/bugfix/hypothesis-driven.md 1.7 KB
- paradigms/dev/architecture-first.md 2.1 KB
- paradigms/enhancement/delta-design.md 2.4 KB
- paradigms/parallel-execution.md 6.0 KB
- phases/_template.md 509 B
- phases/architecture.md 5.7 KB
- phases/code-review.md 7.3 KB
- phases/deliver.md 1.2 KB
- phases/diagnose.md 2.6 KB
- phases/extract-contracts.md 1.8 KB
- phases/impact-analysis.md 2.1 KB
- phases/implement.md 1.8 KB
- phases/provision-environment.md 4.0 KB
- phases/test-plan.md 8.9 KB
- phases/validate-requirements.md 1.4 KB
- phases/verify.md 7.5 KB
- principles/abstraction-layers.md 5.6 KB
- principles/cohesion-separation.md 6.7 KB
- principles/documentation.md 6.4 KB
- principles/error-handling.md 7.1 KB
- principles/information-hiding.md 6.5 KB
- principles/module-depth.md 4.3 KB
- principles/naming-obviousness.md 8.4 KB
- principles/strategic-design.md 10 KB
- README.md 7.2 KB
- references/project-patterns/.gitkeep 0 B
- scripts/match-project.sh 753 B runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 170 lines · 39 tokens per session scan A 91464a38db80
auto-dev is a skill published in the GitHub repository A7um/zero-review (45 stars, last pushed 2mo ago), licensed MIT. It adds 39 tokens to every session and 2,409 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
omni-mcp
Connect to the OmniRoute MCP server (110 tools, 3 transports: SSE/stdio/HTTP). Covers routing, cache, compression, memory, skills, providers, and audit tools across 33 permission scopes.
add-resend
Add Resend (email) channel integration via Chat SDK.
agui-dotnet-sample-step
Add a GettingStarted sample Step (a Server/Client pair) to the AG-UI .NET SDK that demonstrates one protocol feature the way we want users to write it. USE FOR: adding a new samples/GettingStarted/StepNN Server+Client pair, wiring it into AGUI.slnx and the integration-test project, giving it a deterministic…
a-evolve
Apply A-Evolve's agentic evolution methodology to improve AI agent performance across runs. Use when the user wants to diagnose agent failures, generate targeted skills from error patterns, evolve system prompts, or accumulate episodic knowledge. Works standalone or inside AutoResearchClaw pipelines. Triggers on…
workflows
Execute a workflow script that orchestrates multiple subagents deterministically. Use when the user explicitly requests a workflow, multi-agent orchestration, parallel or sequential agent pipelines, structured agent outputs, or a durable background workflow run.
agent-tuning
Use when changing agent model or effort configuration, adapter mappings, or eval candidate profiles.