The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access detected. The skill operates entirely within the local workspace environment.
[COMMAND_EXECUTION]: The skill invokes the ao command-line tool to perform knowledge management tasks such as activate, beliefs, playbooks, and brief. These operations are transparently documented and align with the skill's stated purpose.
[PROMPT_INJECTION]: The skill processes data from a harvest catalog to generate briefings and playbooks, which creates a potential surface for indirect prompt injection if the source data contains malicious instructions. However, this is the primary intended function of the skill, and no specific vulnerabilities were detected.
Ingestion points: .agents/harvest/latest.json and knowledge artifact files.
Boundary markers: Not explicitly defined for the generated output files.
Capability inventory: Command execution via the ao binary for lifecycle management.
Sanitization: Content is processed and consolidated according to the defined DAG; explicit sanitization steps for external content are not described.

knowledge-activation