The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes data from external observability platforms and feature-flag services to verify Gate 4 (Healthy at runtime) and Gate 5 (Reachable by users). This represents an indirect prompt injection surface where untrusted data within logs, metrics, or flag configurations could potentially influence the agent's generated evidence report.
Ingestion points: Monitoring queries (Gate 4) and feature-flag status APIs (Gate 5) ingested via WebFetch or Bash tools.
Boundary markers: The skill does not explicitly define delimiters for external data in its output template.
Capability inventory: The skill has access to Bash, WebFetch, Read, and Glob tools across its verification logic.
Sanitization: No specific sanitization or filtering of the ingested external content is described.
[DATA_EXFILTRATION]: The skill utilizes the WebFetch tool to retrieve status information from external services. While these operations are directed at monitoring platforms and feature-flag providers consistent with the skill's purpose, they represent network activity. No evidence of unauthorized credential or sensitive file exfiltration was found.
[COMMAND_EXECUTION]: The skill employs the Bash tool for verifying system status and checking local deployment environments. These commands are integral to the verification gates described in the process and do not appear to perform unauthorized or privileged operations.

closed-loop-delivery