boshu2

Hard-blocks edits outside declared frozen directories using the PreToolUse hook.

Convert AgentOps skill formats.

Run multi-judge consensus.

Validate OpenShift Grafana dashboards.

Scaffolds a new Claude Workflow script for deterministic multi-agent orchestration. This skill is triggered by phrases like 'build a workflow' or 'create a workflow'.

Run post-implementation validation.

Load relevant .agents context.

Use official OpenAI docs.

Create or refine the PRODUCT.md file.

Research an upstream repo.

Validate PR scope and quality.

Stress-test plans before work.

Probe docs and skills.

Reverse-engineer product specs.

Scaffold a new SKILL.md.

Provides repository coding standards.

Dispatch parallel agents.

Track issues with bd/br, triage with bv, and convert plans to beads.

Initializes AgentOps project files.

Find focused refactor hotspots.

Execute epics through waves.

Run autonomous improvement loops.

Write compact session handoffs.

Repair skill hygiene.

Decompose goals into issue plans.

Generate and validate repo docs (default), READMEs (--mode=readme), and OSS doc packs (--mode=oss).

Canonical vocabulary for human-AI software work.

Use official OpenAI docs.

Run discovery, crank, validation.

Create project, component, or boilerplate scaffolds.

'Audit an existing SKILL.md against the unified AgentOps template (15

Show AgentOps work status.

Dispatch parallel agents.

Trace decisions through artifacts.

Explain AgentOps workflows.

Validate code readiness.

Prepare PR commits and body.

Implement a scoped OSS PR.

Review diffs, find mocks, scan bugs, audit code.

Agent init prompt.

Show AgentOps work status.

Run release validation.

Explore and write findings.

Capture a session learning.

Explain AgentOps workflows.

Compile .agents knowledge wiki.

Run multi-judge consensus.

Load relevant .agents context.