The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions in SKILL.md include example code blocks that demonstrate how to access sensitive system state, specifically through the commands security list-keychains, security find-identity -v, and env | grep IDENTITY. These commands are used to inspect cryptographic identities and environment secrets, which poses a risk of credential exposure.
[COMMAND_EXECUTION]: The script find-polluter.sh executes shell commands via npm test on files discovered through a user-specified search pattern. This functionality allows for the execution of arbitrary code within the context of the test runner.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data such as error messages, logs, and stack traces to identify root causes. Ingestion points: System logs, error messages, and stack traces processed during Phase 1 (Root Cause Investigation). Boundary markers: Absent; there are no instructions to the agent to distinguish between legitimate log data and potential injection attempts within those logs. Capability inventory: The skill utilizes shell execution (via find-polluter.sh) and file reading capabilities. Sanitization: Absent; the skill does not specify any validation or sanitization routines for the external data it processes.

systematic-debugging