Cadence logoCadencev0.2.1
Docs/Agent Skills

Agent Skills

Embed Cadence's AI detection capabilities into agents and automation workflows

Cadence exposes its analysis capabilities through an AI skills system that operates at two levels:

  • Internal skills — four Go-native skills that run as part of the Cadence analysis pipeline, backed by OpenAI or Anthropic
  • External skills — two skills exposed via skills.json for consumption by Claude, ChatGPT, and compatible agent frameworks

How Skills Work

Skills are discrete, purpose-built AI prompts that take structured input (a code snippet, a commit, a detection result, or a full report) and return structured JSON output.

Internally, each skill is executed by the SkillRunner, which:

  1. Formats the input with a system prompt specific to that skill
  2. Calls the configured AI provider (OpenAI or Anthropic)
  3. Parses the structured JSON response
  4. Returns the result alongside any AI-generated fields

Prerequisites

To use skills, AI must be enabled:

Bash
export CADENCE_AI_ENABLED=true
export CADENCE_AI_PROVIDER=openai      # or: anthropic
export CADENCE_AI_KEY=sk-...           # your provider API key
export CADENCE_AI_MODEL=gpt-4o-mini   # optional: override default model
VariableDefaultDescription
CADENCE_AI_ENABLEDfalseMust be true to activate skills
CADENCE_AI_PROVIDERopenaiProvider: openai or anthropic
CADENCE_AI_KEY—API key for the chosen provider
CADENCE_AI_MODELgpt-4o-mini (OpenAI) / claude-3-haiku-20240307 (Anthropic)Model override

Skill Guides

  • Internal Skills — code_analysis, commit_review, pattern_explain, report_summary
  • External Skills — skills.json, Claude tool-use, ChatGPT function calling
BackHome