Adac Skills — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Adac Skills (Plugin) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Languages: English | 中文
ADAC stands for Acceptance-Driven Agentic Coding: a formal delivery mechanism for AI-assisted software engineering. It turns "let AI write code" into a controlled workflow: define context, scope, acceptance evidence, and human gates before expanding implementation work.
Search phrase: Acceptance-Driven Agentic Coding is the full method name behind ADAC Skills.
This repository contains the canonical ADAC specification, reusable delivery templates, portable agent skills, and lightweight validation tooling. It is intended to work across Claude, Codex, and other AI coding agents that can consume SKILL.md-style guidance.
adacadac-skills/
├── .claude-plugin/plugin.json
├── .claude-plugin/marketplace.json
├── .codex-plugin/plugin.json
├── bin/adac-validate
├── docs/
│ ├── brand.md
│ ├── examples.md
│ ├── publishing-checklist.md
│ └── distribution.md
├── README.md
├── examples/
│ └── worker-module.adac.md
├── package.json
├── schemas/
│ └── adac-plan.schema.json
├── scripts/
│ └── validate_adac_plan.py
├── skills/
│ └── adac/
│ ├── SKILL.md
│ ├── agents/openai.yaml
│ └── references/adac-reference.md
└── spec/
├── ADAC-0001.md
└── templates/
├── adac-gate-record.md
└── adac-plan.mdADAC is built on four required controls:
For user-facing or business-visible work, ADAC also asks for User Verification Scenarios: concrete behavior checks a user, reviewer, QA owner, or product owner can run, with preconditions, steps, expected visible results, and evidence to capture.
The mechanism is intentionally stricter than ordinary prompt guidelines and broader than TDD. It prioritizes behavior tests, regression checks, pressure/performance evidence, observability, and rollout safety over coverage theater.
Install with the portable skills CLI:
npx skills add HaloForgeAI/adac-skills --skill adacpython3 scripts/validate_adac_plan.py examples/worker-module.adac.mdThe validator checks required sections, AC status and evidence links, risk-class-specific requirements, user verification scenarios, and vague criteria that lack thresholds or reproducible scenarios.
Or through the npm package entrypoint after publishing/installing:
npx adac-skills validate examples/worker-module.adac.mdADAC Skills is prepared for several distribution paths:
HaloForgeAI/adac-skillsadac-skills.claude-plugin/plugin.json.claude-plugin/marketplace.json.codex-plugin/plugin.jsonskills/adac/See brand notes, example workflows, publishing checklist, and distribution notes for naming, usage, publishing, and registration options.
The static project page lives in docs/index.html. Enable GitHub Pages from the main branch and /docs folder to publish it.
The repo structure borrows from several formal mechanisms:
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.