The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform various diagnostic tasks, such as running tests, CLI tools, and the provided scripts/hitl-loop.template.sh for interactive debugging.
[DYNAMIC_EXECUTION]: The diagnostic workflow involves creating and running local scripts (e.g., reproduction harnesses, instrumentation scripts) to verify bug behavior and test hypotheses. This is a core component of the feedback loop methodology described in Phase 1 and Phase 4.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it is designed to ingest and analyze external data such as application logs, captured network traces, and error reports.
Ingestion points: The skill processes external content through log files, replayed traces, and user-provided error messages captured by the hitl-loop.template.sh script.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the ingested logs or traces.
Capability inventory: The agent has access to powerful tools including Bash, Write, Edit, and Read, which could be targeted by instructions hidden in processed data.
Sanitization: The skill does not implement specific sanitization or filtering for the external diagnostic data it consumes.

solo-diagnose