The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/measure.py executes local system commands to interact with Git. It calls git rev-parse --show-toplevel to identify the repository root and git check-ignore to respect .gitignore rules when scanning the project tree. These calls are implemented using subprocess.run with static arguments and no shell invocation.- [DYNAMIC_CONTEXT_INJECTION]: The skill's instruction file SKILL.md uses the !command syntax to provide the agent with project context at load time. It executes the local measure.py script to calculate instruction file sizes, runs git status to identify modified instruction files, and performs checks for standard documentation files like README.md and CONTRIBUTING.md.- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from the repository it is executed against, which creates a surface for potential injection attacks. * Ingestion points: The skill reads AGENTS.md, CLAUDE.md, and README.md files from the target repository. * Boundary markers: No explicit delimiters or boundary markers are defined to isolate untrusted content from the agent's logic. * Capability inventory: The skill allows the agent to read, write, and delete markdown files and execute Git commands. * Sanitization: Content from project files is directly processed to build a rewriting ledger without evidence of sanitization or instruction filtering.

normalize-agents-md