The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill reads the .env file to extract and verify port numbers and environment variable names. Accessing .env is a security concern as it typically contains sensitive application secrets and credentials.
[COMMAND_EXECUTION]: The skill executes multiple shell commands such as find, grep, git, and ls to gather information about the project structure, configuration, and history.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted documentation files (e.g., README.md, ARCHITECTURE.md) and configuration files. Malicious instructions placed in these files could potentially influence the agent's actions, particularly given its ability to modify files and commit changes to the repository.
Ingestion points: Reads content from project documentation (*.md) and configuration files (docker-compose.yml, .env).
Boundary markers: No explicit delimiters or instructions are used to separate data from potentially malicious instructions within these files.
Capability inventory: File system modification, shell command execution, and repository state changes via git add and git commit.
Sanitization: No sanitization or validation is performed on the content of the audited files before processing.
[DYNAMIC_EXECUTION]: The skill attempts to execute a local reconnaissance script (scripts/recon.sh) if it is present within the skill's plugin directory.

doc-audit