debugging-toolkit-smart-debug
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input through the
$ARGUMENTSplaceholder, which is used to provide the context for debugging tasks. - Ingestion points: Untrusted data enters the agent context through the
$ARGUMENTSvariable referenced in the 'Context' section and at the end of theSKILL.mdfile. - Boundary markers: The skill lacks delimiters (such as XML tags or unique markers) or specific instructions to ignore embedded commands within the ingested data, increasing the risk of the agent following instructions hidden inside error reports or stack traces.
- Capability inventory: The skill uses a
Task tool(subagent), observability platform connectors (Sentry, DataDog, New Relic), and includes steps for fix generation and validation (code execution). - Sanitization: There is no evidence of input sanitization or validation for the data provided via
$ARGUMENTS.
Audit Metadata