agentic-workflow-framework

Plugin agentic-workflow-framework from marketplace agentic-workflow-framework in jeanfbrito/agentic-workflow-framework. Commands read from the repo's manifest — yesterday.

Skills (4)

• agentic-workflow

  Manage task ledgers, handoffs, and authorized delegation for work spanning sessions or agents.

• harness-leak-review

  Finds and removes personal harness tools in the guidance of a shared repository. Personal tools are your MCP servers, plugins, agents and skills, ~/.claude paths and agentic-workflow ledgers. The goal is guidance that works for a teammate who has none of them. Use it to review or clean the agent gui

• personal-engineering-rules

  Apply Jean's conventions. Use when porting reference behavior, fixing a bug, writing or keeping a test, shaping a module boundary, evaluating review feedback, or writing engineering reports.

• skill-review

  Runs the weekly skill review. Measures which skills, commands and plugin skills actually ran in local Claude Code sessions. Counts wiki pages, missed wikis and missed skills. Sorts each skill into keep, merge, promote or retire, and applies only the approved changes. Also sets the user's wiki folder

Slash commands (6)

• /agentic-workflow-framework:agentic

  Carry a task through completion with proportionate planning and project continuity.

• /agentic-workflow-framework:blocker

  Append a decision blocker to .localdev/workflow/blockers.md in canonical format, with the default action to take if no answer arrives.

• /agentic-workflow-framework:handoff

  Write a cross-session handoff for the current task, so the next session can resume with full context

• /agentic-workflow-framework:init-agentic

  Initialize project task ledgers and known-issue records while preserving existing content.

• /agentic-workflow-framework:known-issue

  Document a persistent platform or dependency constraint in docs/KNOWN_ISSUES.md — the kind of thing the next developer or agent would re-discover otherwise

• /agentic-workflow-framework:qq

  Quick question answered by haiku — one fast turn, session model untouched

Subagents (10)

• auditor

  Diagnose the root constraint after two failed approaches and return an evidence-based revised brief.

• builder-fast

  Implement a clear, scoped feature or fix and prove its focused completion criteria.

• builder-smart

  Implement reasoning-heavy or concurrency changes, fix a failed builder-fast attempt, or write prose for a public repo.

• builder-trivial

  Apply one fully specified repetitive transformation across assigned targets.

• finder

  Locate code and trace call chains for a bounded, read-only question.

• planner

  Plan ambiguous architecture or risky changes with a scoped brief and focused completion criteria.

• researcher

  Find primary-source evidence for a specified API, library, CLI, or platform question.

• reviewer

  Review a risky diff for concrete regressions and scope issues, using existing verification evidence.

• tester

  Validate integration boundaries, fill missing proof, or perform explicitly requested independent checks.

• watcher

  Run a finite noisy job or digest an existing log, returning the verdict and exact errors.

Hooks (7)

• SessionStart

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/inject-agentic-context.sh"

• SessionStart

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/session-scan.sh"

• SessionStart

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/skill-review-nudge.sh"

• PreCompact

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/pre-compact.sh"

• Stop

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/stop-ledger-audit.sh"

• UserPromptSubmit

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/orchestrator.sh"

• PreToolUse

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/allow-workflow-paths.sh"

Similar plugins