The Agent Skills Directory

[SAFE]: The skill consists of instructional markdown files that guide an AI agent through a systematic debugging workflow. It includes specific guardrails to prevent common security pitfalls in autonomous agents. The instructions in SKILL.md and references/probe-escalation.md explicitly forbid the exposure of secrets or user data during instrumentation and logging. The skill mandates verification steps (artifact freshness, self-review) to ensure that repairs are correct and do not introduce regressions. The 'Delegated Diagnosis' section provides safe guidelines for using sub-agents or host orchestration tools, focusing on model capability and cost-effectiveness rather than arbitrary code execution. While the skill processes user bug reports (untrusted input), it contains internal logic to treat this input as evidence to be verified rather than commands to be followed blindly, mitigating potential injection risks.
[NO_CODE]: The skill is entirely composed of markdown documentation and instructions for agent behavior, containing no executable scripts, configuration files that trigger code execution, or external package dependencies. The instructions for 'delegated diagnosis' and 'probes' refer to platform-level capabilities or temporary debugging steps that the agent must manage safely according to the provided rules.

vibe-debug