Skills Explorer Add Skills

constraint-driven-development

An agent skill by addyosmani, from addyosmani/agent-skills. Tags: debugging, developer-tools, documents.

What it does

Establishes a project's quality bar as a written contract and stops agents quietly lowering it. Interviews the user on which dimensions matter, supplies sane default thresholds when they have no number in mind, records everything in CONSTRAINTS.md, and watches the diff for a weakened bar — new @ts-ignore or eslint-disable suppressions, skipped or deleted tests, assertions stripped out, unimplemented stubs, thresholds edited down. Use when no quality bar is written down, when the user says "set up constraints" or "define our standards", when the user wants dimensions they care about — accessibility, web performance, coverage — set up as enforced constraints, when an agent keeps silencing checks or skipping tests to get to green, when you need a coverage or performance threshold and don't know what number to pick, or when an agent writes more code than anyone will read.

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add addyosmani/agent-skills --skill constraint-driven-development

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/addyosmani/agent-skills
cp -r agent-skills/skills/constraint-driven-development ~/.claude/skills/constraint-driven-development

Safety box score

Grade C, moderate risk: 37 of 100 risk points, rated 2026-10-21. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill mainly sets up project-wide quality gates by creating `CONSTRAINTS.md`, updating agent guidance files, and wiring checks like tests, lint, and scanners. The main risks are local environment changes from installing tools, lots of command execution, and the possibility of leaking secret-like matches into logs if redaction is mishandled.

Category levels, from 0 (none) to 3 (high)
CategoryLevelWhy
Code & shell execution2 (moderate)It is built around running real commands such as `tsc --noEmit`, `vitest run --coverage`, `semgrep scan`, and a floor guard that shells out to `git` via `execFileSync('git', args...)`.
Network access2 (moderate)It depends on package downloads and live URL checks, e.g. `brew install gitleaks`, `npm i -D lighthouse`, and `axe $URL` / `lighthouse $URL`, so setup and some checks require network access.
Filesystem reach2 (moderate)It writes a new root-level `CONSTRAINTS.md` and tells the agent to modify `AGENTS.md` and `CLAUDE.md`, which changes repo-wide behavior and review rules.
Secrets & credentials1 (low)It adds secret scanning with `gitleaks detect --redact --no-banner` and explicitly warns that unredacted matches can land in transcripts, but it does not ask for secret-store access.
Privilege & persistence2 (moderate)The setup includes package installs like `brew install gitleaks`, `brew install osv-scanner`, `pip install mypy`, and `npm i -D lighthouse`, which modify the local user environment.
Instruction hijack surface0 (none)It does not instruct the agent to ignore higher-priority rules, follow remote instructions, or hide actions; the only meta text is normal process guidance and a red flag about `--no-verify`.
Irreversible actions0 (none)It does not deploy, publish, merge, pay, book, or delete anything; it only defines and enforces quality checks.
Transparency0 (none)All shipped content is plain markdown, including the reference guard code, with no binaries, hidden payloads, or unpinned remote installers.

Before installing

Rated 2026-10-21 by an automated review of 2 files from the skill's repository.

Source

Repository
addyosmani/agent-skills (all skills from this repository)
Path
skills/constraint-driven-development/SKILL.md
Branch
main
Updated
2026-09-19