Curated skills, commands, and agents for AI-assisted product management (and life)
Agent OpenCode
Independently re-fetches all sources cited in a draft (Jira tickets, GitLab files, wiki pages, doc URLs, MRs) and verifies that each exists and supports the claims made. Returns a structured verification report. Use as a validation step in any command that produces source-backed analysis.
Agent OpenCode
Adversarial reasoning reviewer — finds logic errors, weak explanations, unsupported claims, and gaps that mechanical checklists miss. Read-only — reports issues, never fixes them.
Agent OpenCode
Reviews code for bugs, security, and maintainability with tool-assisted validation.
Agent OpenCode
Ruthless style editor for work and personal writing. Detects work-vs-personal context from the file path, loads the right style guide, and runs ten scoped passes to weed out AI patterns. Read-only — reports findings with file:line and suggested rewrites, never fixes them.
Agent OpenCode
Assesses whether a Jira ticket's fix can be implemented by a technical PM and an LLM coding agent. Reads GitLab code, maps change surface area, scores feasibility criteria. Returns structured evidence for the orchestrator to make a Go/No-Go verdict.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: