The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform shell operations within the user's environment to fulfill its primary purpose. This includes running project-level tasks such as npm run build and npm test, as well as executing a local script (node render.mjs) for generating visual reports.
[INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves reading and analyzing untrusted source code from the project directory, creating a significant attack surface for indirect prompt injection.
Ingestion points: The agent reads arbitrary source files during the 'Diagnose' phase and the 'Knowledge Import' scan to identify algorithmic patterns.
Boundary markers: There are no instructions defining delimiters or safety boundaries for the ingested code, meaning the agent may follow malicious instructions embedded within the analyzed files.
Capability inventory: The agent has broad capabilities including rewriting project source code, creating a new .algo-profile/ directory with markdown documentation, and executing shell commands (npm, node).
Sanitization: The instructions provide no guidance for sanitizing or escaping content read from files before it is processed by the LLM.

data-algo