The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a Bash script (scripts/abridge-quick-diag.sh) that executes network diagnostic commands including curl, openssl, and dig. These commands are used to verify API health, TLS certificate validity, and DNS resolution within the user's environment.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API endpoints (e.g., /encounters/sessions) and processes application logs to generate debug bundles. This external data represents a potential attack surface if the retrieved content contains instructions designed to influence the agent when it subsequently reads or processes the diagnostic output.
Ingestion points: Data is retrieved via the collectSessionDiagnostics function in SKILL.md and from local application logs.
Boundary markers: The skill does not employ specific delimiters or "ignore" instructions when writing the collected data to the JSON bundle.
Capability inventory: The skill performs file writes (fs.writeFileSync), network requests (fetch, curl), and shell command execution.
Sanitization: The skill includes a redactPhi function that uses regular expressions to identify and mask SSNs, MRNs, patient names, and dates of birth before the data is saved to disk.

abridge-debug-bundle