The Agent Skills Directory

[PROMPT_INJECTION]: The skill workflow involves analyzing untrusted external data, such as log dumps, HAR files, and network traces, to diagnose issues. This presents an indirect prompt injection surface where malicious content within these files could attempt to manipulate the agent's logic.
Ingestion points: Phase 1 of SKILL.md identifies logs, HAR files, and core dumps as artifacts to be analyzed.
Boundary markers: There are no instructions to use specific delimiters or protective prompts when reading these external files.
Capability inventory: The agent has the ability to execute shell commands, run tests, and use curl across various phases of the debugging loop.
Sanitization: No sanitization or validation of the content of the analyzed artifacts is mentioned.
[COMMAND_EXECUTION]: The skill facilitates the creation and execution of local reproduction scripts, including the use of a provided bash template scripts/hitl-loop.template.sh. It also makes use of common CLI tools such as curl and git for automated bisection and testing. These actions are within the expected scope of a software development and debugging skill.

diagnose