Plugin Claude Code
RoT DTD GOAL - a /goal engine where completion is earned by re-running real acceptance criteria, not declared by the model. Simultaneous verification, clock-free flake scoping, a fenced trust boundary, a dependency-ordered goal queue, and a Lean 4 specification the kernel re-verifies.
Plugin Claude Code
A /goal engine whose completion is EARNED, not announced. Acceptance criteria are real shell commands; the Stop hook re-runs every one of them and only exits 0 lets a goal finish. Verification defends itself: a sealed integrity ledger re-hashes each criterion before the gate trusts it, a negative-control red team…
Agent
Reads the cross-goal record. Invoke when the same kind of goal keeps stalling, exhausting its budget, or escalating — reconstructs the recurring pattern from history.tsv and past journals across many goals, without predicting the next one will succeed.
Agent
Audits the trust contract against the code. Invoke after changing anything the engine prints, adding a record field, or adding an agent — finds strings the engine says that the contract never declared.
Agent
Breaks stalls. Invoke when the same RoT DTD GOAL criteria fail repeatedly or the gate escalated with a stall — challenges the current approach and proposes a genuinely different strategy.
Agent
Reads the record. Invoke when a goal escalated, stalled, or completed in a way nobody understands — reconstructs what actually happened from the journal, ledger, timings and seal generations, without guessing.
Agent
Decomposes a user goal into 3-10 machine-checkable acceptance criteria with real single-line shell verify commands. Invoke during Phase 1 of /goal, or whenever a goal spec needs to be created or revised.
Agent
Decomposes a large goal into a dependency-ordered queue of smaller goals. Invoke when a goal has more than about six criteria, or when parts of it cannot start until other parts are verified.
Agent
Attacks an acceptance criterion. Invoke when a criterion looks too easy, survived the empty-directory control, or passed while its files were damaged — finds the input that makes it fail, or proves it cannot.
Agent
Read-only auditor that runs the RoT DTD GOAL verification suite and diagnoses failures without modifying anything. Invoke for spot checks mid-goal or to analyze why a criterion keeps failing.
Command
Abort the active goal (state kept on disk for post-mortem).
Command
Dispatch ONE of the DTD-declared agents on a task — agent name validated against the trust contract, model and effort selectable per call.
Command
Attack the goal's own acceptance criteria — ledger integrity, negative control, and mutation probe (does not consume an iteration).
Command
Pause the active goal — the stop-gate goes dormant until resumed.
Command
Resume a paused or human-escalated goal (resets iteration budget and stall detector).
Command
Show the RoT DTD GOAL goal, criteria pass/fail, iteration budget, and journal tail.
Command
Fan out ALL the DTD-declared agents (or an agents= subset) on one thing at once — a workflow when the harness has one, parallel dispatches otherwise — model and effort selectable.
Command
Run all RoT DTD GOAL acceptance criteria right now and report (does not consume an iteration).
Command
Pursue a goal to verified completion via the RoT DTD GOAL state machine (verification-owned, no promise strings, no blind loops).
Hook
Runs when the agent finishes a response, executing stop_gate.sh via bash. From Nova-Violet-Role/RoT-DTD-GOAL.
Hook
Runs on the StopFailure event, executing journal_event.sh via bash. From Nova-Violet-Role/RoT-DTD-GOAL.
Hook
Runs when a session starts, executing context.sh via bash. From Nova-Violet-Role/RoT-DTD-GOAL.
Hook
Runs when a session ends, executing journal_event.sh via bash. From Nova-Violet-Role/RoT-DTD-GOAL.
Hook
Runs once when the project is set up, executing journal_event.sh via bash. From Nova-Violet-Role/RoT-DTD-GOAL.