Audit specs, plans, code, and content against evidence with YAML verdict artifacts.
npx skills add https://github.com/michael-conrad/.opencode --skill audit-michael-conrad
Please help me install this Agent Skill. Skill: audit Source: https://github.com/michael-conrad/.opencode/tree/main/skills/audit Command: npx skills add https://github.com/michael-conrad/.opencode --skill audit-michael-conrad
The Audit skill provides a structured approach to verify accuracy, coherence, and fidelity of specs, plans, code, and generated content, ensuring claims are supported by evidence and drift is detected early.
Invoke the audit workflow and dispatch a verification task to a clean-room sub-agent using task(subagent_type="general").
High-intent search queries and answers about installing and using this skill.
Spec fidelity verification and drift detection are handled by dispatching clean-room sub-agents to cross-check content against evidence. This process enforces live-tool validation and generates YAML verdict artifacts to ensure claims are fully supported and auditable.
A clean-room audit workflow isolates verification logic by dispatching tasks to sub-agents, shielding the orchestrator from the verification work. This ensures objective cross-validation of specs, plans, and code with structured evidence trails.
Auditing generated code and plans involves dispatching a verification task to a clean-room sub-agent. The sub-agent cross-checks the output against original specs, detects any coherence drift, and writes YAML verdict artifacts for downstream processes.
Yes, guideline audits are dispatched to clean-room sub-agents via task(subagent_type="general"). This shields the orchestrator from verification work while enforcing live-tool evidence and generating reproducible YAML verdict artifacts.
YAML verdict artifacts are required to feed downstream cross-validation and completion processes. They provide a structured, reproducible evidence trail that ensures auditable results when verifying content, specs, or code fidelity.
Yes, the audit workflow supports plan fidelity checks and cross-validation across teams. It enforces live-tool evidence and generates YAML verdict artifacts to ensure ongoing alignment and detect spec drift effectively.
michael-conradRoute planned work to clean-room sub-agents for isolated step execution.
GaneshMadarasuDetect data and concept drift in Kafka anomaly scores using ADWIN and DDM detectors.
michael-conradValidate success criteria and collect artifacts before task completion.
dnhessAudit agent work for intent, consistency, constraints, and trajectory coherence.
