The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests content from various local files to generate project-specific rules, creating a potential vector for indirect prompt injection if those files contain malicious instructions.
Ingestion points: The agent is instructed to gather evidence from files such as package.json, README.md, existing AGENTS.md, lockfiles (bun.lock, pnpm-lock.yaml, etc.), and .env.example in the local repository.
Boundary markers: The instructions do not specify the use of clear delimiters or explicit warnings to the agent to ignore instructions embedded within the source files when generating the new AGENTS.md or documentation.
Capability inventory: The skill allows for the creation and modification of files (AGENTS.md) and the generation of a directory structure and markdown files within docs/.
Sanitization: No explicit sanitization or validation logic is provided to filter content extracted from local repository files before interpolation into the generated documentation.
[COMMAND_EXECUTION]: The skill relies on shell commands to inspect the repository environment and gather context for bootstrapping.
Evidence: The workflow includes the execution of rg (ripgrep) to search for configuration files and find to map the directory structure of the docs/ folder.

project-bootstrap