The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to create and run various feedback loops, such as unit tests, fuzzing scripts, and headless browser interactions. This is a primary function of the skill's purpose (diagnosis and debugging) and follows standard engineering practices.
[COMMAND_EXECUTION]: The workflow incorporates the use of CLI tools, version control (git bisect), and network utilities (curl, wget) to automate the reproduction of reported bugs.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data from bug reports, log files, and network traces. However, this is inherent to the debugging task and the skill does not grant excessive privileges or perform sensitive operations based on that data without context.
Ingestion points: User bug reports, log files, network traces, and DOM snapshots mentioned in Phase 1 of SKILL.md.
Boundary markers: None specified in the instructions.
Capability inventory: The agent is permitted to write scripts, perform network requests, and execute test runners as described in SKILL.md and scripts/hitl-loop.template.sh.
Sanitization: No explicit sanitization or filtering of external logs or traces is specified.
[SAFE]: The provided Bash template (scripts/hitl-loop.template.sh) is a benign helper for terminal-based user interaction during reproduction steps.

diagnose