The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified. The skill is designed for informational purposes and lacks the capability to modify the environment.
[COMMAND_EXECUTION]: The skill invokes vendor-specific commands (de-novo skills understory reading and de-novo skills mycelium query) to retrieve project metadata. These are recognized as legitimate functions of the de-novo platform for status reporting.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from other tools, which is a potential surface for indirect injection. Ingestion points: Data enters the context via the output of understory and mycelium tools referenced in SKILL.md. Boundary markers: The instructions include explicit constraints to 'say nothing they do not hold' and to name guesses as such, which limits the agent's reliance on potentially injected data. Capability inventory: The skill is limited to reading and speaking; it is explicitly prohibited from writing files, proposing assertions, or modifying the document state. Sanitization: The platform-level disable-model-invocation: true configuration in SKILL.md prevents the agent from triggering further autonomous actions based on the ingested content.

clearing