The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses sensitive configuration via the .env file and environment variables (FOUNDRY_TOKEN, FOUNDRY_CLIENT_SECRET). It includes proactive redaction logic using sed to mask values in the collected configuration and only reports the presence/length of tokens in the summary report.
[DATA_EXFILTRATION]: Performs an API connectivity test using curl to a user-defined hostname (FOUNDRY_HOSTNAME) including the FOUNDRY_TOKEN in the Authorization header. This is a legitimate diagnostic function for the service.
[COMMAND_EXECUTION]: Executes shell commands via bash to gather system information, check installed Python packages, and search log files for specific error patterns.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local *.log files to extract error context.
Ingestion points: Reads contents of *.log and .env files using grep and sed.
Boundary markers: None explicitly present for log reading, though data is redirected to a temporary file.
Capability inventory: Uses curl, tar, and standard shell utilities (grep, sed, mkdir).
Sanitization: Implements sed-based redaction of variable values in the .env file before including them in the diagnostic bundle.

palantir-debug-bundle