The Agent Skills Directory

[DATA_EXFILTRATION]: The skill explicitly directs the agent to locate and access highly sensitive configuration and credential files within the project environment.
Evidence: The instructions mandate that the agent "thoroughly explore the codebase" and specifically "Find the... Credentials setup (config/credentials.yml.enc, config/master.key)".
Evidence: The skill requires identifying environment variables from files such as .env.example or .env.sample, which may contain sensitive context.
[INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and processing content from an untrusted codebase, which is a primary vector for indirect prompt injection attacks.
Ingestion points: The agent is instructed to read the root directory structure, database schemas (db/schema.rb), configuration files, and key dependency lists (Gemfile, package.json).
Boundary markers: There are no instructions to treat codebase content as untrusted or to use delimiters to prevent the agent from following instructions embedded within the files being analyzed.
Capability inventory: The agent has the capability to write the resulting content to the README.md file in the project root.
Sanitization: The skill lacks any requirement to sanitize, escape, or validate the information extracted from the project files before it is used to generate the final documentation.

readme