launchdarkly-automations-cleanup-settings
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages a configuration surface that ingests untrusted data which is later interpreted as instructions by a downstream agent.
- Ingestion points: The skill reads settings from the LaunchDarkly API via the
get-automationstool and accepts user-provided input for thecustomContextfield as described inSKILL.md. - Boundary markers: There are no specified delimiters or boundary markers defined in the instructions to isolate these custom instructions from the rest of the agent's context during interpolation.
- Capability inventory: The skill uses the
update-automationstool to persist these instructions, which then govern the behavior of an automated cleanup agent capable of opening pull requests in GitHub. - Sanitization: The skill lacks explicit sanitization or validation logic for the
customContextfield, relying instead on manual user confirmation before applying changes.
Audit Metadata