The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the eval command within scripts/cp2-verify-build.sh to dynamically run build and test commands (such as npm test, pytest, or cargo test) based on the detected project environment. This is a primary feature designed to ensure code quality through automated verification.
[REMOTE_CODE_EXECUTION]: The installation process involves downloading a script from the author's GitHub repository (miles990/self-evolving-agent) and piping it directly to the shell. Additionally, the scripts/setup-skill-index.sh script downloads and executes the uv package manager installer from astral.sh, which is a well-known technology service.
[EXTERNAL_DOWNLOADS]: The skill uses scripts/sync-skills.sh to download and index skill metadata from other repositories owned by the author (claude-software-skills and claude-domain-skills) to facilitate skill recommendations.
[PROMPT_INJECTION]: The skill processes project files and memory logs, creating a potential surface for indirect prompt injection.
Ingestion points: Project source code files, .claude/memory/ records, and user goals in the skills/01-core/_base/goal-analysis.md file.
Boundary markers: The skill employs Markdown headers, checklists, and YAML frontmatter to delineate instructions from external data.
Capability inventory: The skill possesses extensive capabilities including shell command execution, file system modification, and network operations via integrated tools.
Sanitization: Content from the ingested files is used directly within the agent's context without explicit sanitization or filtering logic.

evolve