The Agent Skills Directory

[SAFE]: The skill implements a read-only policy for source code and wiki entries during analysis, ensuring the integrity of the input data.
[SAFE]: Data persistence is confined to specific project subdirectories (.workflow/knowhow/, .workflow/specs/, .workflow/issues/), preventing arbitrary file writes.
[SAFE]: A confirmation gate using request_user_input is required before the agent can create knowledge-gap issues, providing human oversight for automated actions.
[COMMAND_EXECUTION]: The skill invokes several external CLI utilities (e.g., maestro, wiki-connect, manage-wiki) via the Bash tool to perform its analysis. These tools are used for knowledge graph search and health checks consistent with the skill's purpose.
[SAFE]: No evidence of prompt injection, data exfiltration, obfuscation, or unauthorized remote code execution was detected.

wiki-digest