cdli.ai / agent skill collection
Skills agents can read.
Six engineering methods, packaged as SKILL.md folders. Humans can inspect them. Agents can crawl them.
public methods
06 skills
v0.1
index: https://cdli.ai/skills/*/SKILL.md
Install
Clone once. Copy the folders where your agent reads skills.
- 01
clone
git clone https://github.com/cdliai/cdli-agent-skills.gitGets the public collection.
- 02
Claude Code
mkdir -p ~/.claude/skills && cp -R cdli-agent-skills/*/ ~/.claude/skills/Personal skills. Use .claude/skills/ for project-local installs.
- 03
Codex / local agents
mkdir -p ~/.codex/skills && cp -R cdli-agent-skills/*/ ~/.codex/skills/Same SKILL.md folders; place them in your agent's skills directory.
Index
Each row has an HTML page, raw markdown, and public source.
- 01
./reviewing-codeReviews pull requests and diffs for correctness, logic errors, performance, style, and merge risk.
v0.1open ->rawview source - 02
./auditing-securityAudits exposed code and configs for vulnerabilities, injection, auth, secrets, and OWASP risks.
v0.1open ->rawview source - 03
./auditing-tech-debtFinds duplication, dead code, drift, naming decay, complexity, and high-interest technical debt.
v0.1open ->rawview source - 04
./refactoring-codeRestructures code for readability, cohesion, and lower complexity without changing behavior.
v0.1open ->rawview source - 05
./mapping-architectureMaps entry points, data flow, dependencies, and failure modes before architecture decisions.
v0.1open ->rawview source - 06
./reviewing-api-contractsReviews REST and GraphQL contracts for endpoint shape, schemas, versioning, auth, errors, and DX.
v0.1open ->rawview source