The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local Node.js script, spec-superflow.mjs, by replacing the placeholder SSF in command strings. These commands are used to record debug attempts, show attempt ledgers, and record architectural escalation decisions.
[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze potentially untrusted external data, creating an attack surface where instructions could be embedded in the data being processed.
Ingestion points: The agent is directed to read error messages, stack traces, git diffs, and codebase files to investigate root causes (Phase 1) and record failure logs as evidence (DP-5).
Boundary markers: The instructions do not specify the use of delimiters or provide explicit warnings for the agent to ignore instructions found within the logs or code files being analyzed.
Capability inventory: The agent has the capability to execute shell commands via the SSF wrapper script and write failure logs to the local file system within the <change-dir>.
Sanitization: There is no mention of sanitization or validation protocols for the external data before it is processed by the agent.

bug-investigator