The Agent Skills Directory

[SAFE]: Functional Assessment
The skill is designed for administrative maintenance of a markdown-based vault.
The ROLE.md file establishes a restrictive 'secretary' role, enforcing a policy of least privilege and requiring explicit human 'grants' for state changes, which acts as a safeguard against autonomous misuse.
[PROMPT_INJECTION]: Indirect Surface Analysis
Ingestion points: The agent reads 'area notes' in SKILL.md to evaluate standards.
Boundary markers: Not explicitly defined for the untrusted content within notes, but the agent's restricted stance mitigates risk.
Capability inventory: Limited to vault note modification; no network or shell execution capabilities are present.
Sanitization: Not present, but the risk is assessed as safe given the primary purpose and lack of high-risk tools.
[DATA_EXFILTRATION]: No data exfiltration or credential harvesting patterns were found. The skill relies on local configuration variables for paths.

maintain-area