The Agent Skills Directory

[DATA_EXFILTRATION]: Accesses system logs and configuration files.
Evidence: The skill reads ~/.npm/_logs/*.log and .env to extract diagnostic data and environment variables.
Context: A redaction step using sed is implemented to mask environment variable values before they are added to the diagnostic bundle to mitigate the risk of accidental secret exposure.
[EXTERNAL_DOWNLOADS]: Performs a network connectivity test to a remote service.
Evidence: Uses curl to probe the health status of api.coderabbit.com.
Context: The request targets a well-known service domain associated with the skill's primary troubleshooting purpose.
[PROMPT_INJECTION]: Vulnerability surface identified for indirect prompt injection via the processing of untrusted local data.
Ingestion points: Reads content from local application logs (~/.npm/_logs/*.log) and configuration files (.env).
Boundary markers: No delimiters or explicit instructions are provided to the agent to treat the collected content as untrusted data.
Capability inventory: The skill has the capability to read files, create directories, execute shell commands, and perform network connectivity tests.
Sanitization: Employs regex-based redaction to mask secrets within environment variables, though logs remain unvalidated.

coderabbit-debug-bundle