The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/intake_cached.js and its sub-scripts use child_process.spawnSync to execute git and node commands. This is used to gather repository metadata (commit hashes, remote URLs, file lists) and execute internal scanning logic.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing untrusted data from the repository being scanned.
Ingestion points: The tool ingests the repository file tree and the contents of various manifest files such as package.json, go.mod, and Cargo.toml.
Boundary markers: The output provided to the agent lacks explicit delimiters or instructions to ignore embedded commands within the ingested data.
Capability inventory: The skill is configured with powerful tools including Bash, Read, Write, Grep, and Glob in SKILL.md.
Sanitization: While internal cache paths use a slugification function, the metadata and project overview presented to the agent are not sanitized to prevent malicious content from influencing agent behavior.

repo-intake