observability-first-debugging
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill encourages the agent to print environment variables and full API headers/bodies to diagnose issues. This practice can lead to the accidental exposure of sensitive credentials, such as Bearer tokens, API keys, or internal configuration secrets, into the conversation history or logs.
- [INDIRECT_PROMPT_INJECTION]: The skill provides a methodology that involves capturing and displaying output from external systems (APIs, logs, file reads). This creates a surface where untrusted data could be interpolated into the agent's context without sanitization.
- Ingestion points: External API responses, file contents, and environment data processed in SKILL.md examples.
- Boundary markers: None; the skill does not recommend using delimiters or instructions to ignore embedded commands in the data being debugged.
- Capability inventory: Uses print statements and shell commands (set -x, echo) to bring external data into the agent's view.
- Sanitization: None; the methodology does not specify filtering or escaping of the observed data.
Audit Metadata