The Agent Skills Directory

[DATA_EXFILTRATION]: The skill is designed to read the content of local markdown documents and upload them to the Proof web API at www.proofeditor.ai. This behavior is the central purpose of the skill, facilitating document sharing and collaborative review.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute curl for API interactions, jq for data processing, and filesystem commands like mv and mktemp to perform atomic updates to local markdown files during the synchronization process.
[PROMPT_INJECTION]: The Human-in-the-Loop (HITL) review mode introduces a surface for indirect prompt injection by processing untrusted data from an external source.
Ingestion points: Feedback is ingested from the Proof Web API via the GET /api/agent/{slug}/state endpoint (referenced in references/hitl-review.md).
Boundary markers: The skill does not implement explicit boundary markers or delimiters when interpreting document comments or suggestions.
Capability inventory: The agent possesses the capability to modify the local filesystem (Write tool, mv command) and execute network requests (WebFetch, curl).
Sanitization: There is no explicit sanitization logic; the agent is instructed to "decide how to respond" and is encouraged to apply fixes directly based on its interpretation of the feedback.

proof