nuclear-grade plugin

Records the version everyone agreed is correct, the evidence behind it, and what would make that record out of date.

Splits scope into a product-first work breakdown that follows the 100% rule, keeps pieces from overlapping, uses outline numbers, and gives every piece a dictionary entry.

Picks Quick, Standard, or a stronger human-reviewed mode based on consequence, how easy it is to undo, and how much is unknown.

Brings an abandoned or half-filled change packet to an honest terminal state by completing it, closing it with a recorded rationale, or deleting it.

Reviews a diff or module for slipping standards, favoring deletion over rearranging, and ends in one honest verdict.

Prepares focused context for an AI agent, reviewer, verifier, or releaser at the start of a task, with a clear role, goal anchor, authority, evidence to produce, forbidden actions, and stop conditions.

Decides who holds authority for a change \u2014 the agent at the edge or a human gate \u2014 by matching decision rights to reversibility, evidence, and consequence, and names the escalation trigger.

Keeps a standing register of known deficiencies so flaky tests, noisy alerts, unowned services, and recurring incidents get aged, owned, and fixed or formally risk-accepted instead of quietly normalized.

Tests the current work against a written, lasting mission anchor and forces one choice, re-anchor, escalate, or stop.

Checks factual assertions (numbers, quotes, dates, versions, API or tool behavior, what a source actually says) against primary sources read this pass, confirming each load-bearing fact by a second independent means.

Designs a clean folder and file layout as real architecture, building it from a work breakdown or an existing tree, grouping by what changes together and what happens to it, with platform-safe sortable names and a short note per folder.

Screens what a controlled change might leave out of date across docs, tests, skills, commands, templates, checkers, prompts, releases, saved versions, and evidence, and names the triggers to re-check.

Runs a live incident the stabilize-first way \u2014 name a commander, separate facts from hypotheses, prefer reversible actions, communicate on a cadence, and drive corrective actions to closure.

States what an agent intends to do and why before a critical or irreversible action, so a reviewer can challenge the thinking and not just the result.

Turns incidents, near misses, bad handoffs, review surprises, escaped bugs, and signals from real use into lasting fixes to your safeguards.

Reviews public text for license, warranty, compliance, safety, security, certification, and fitness claims that go too far, then rewrites them to stay inside the real limits.

Creates or updates Quick or Standard change records, adds the required files, and refreshes what each claim must prove for an evidence-backed pull request.

Convenes several reviewers on one frozen evidence packet, names the diversity axis each seat actually varies, preserves dissent, and verifies every finding against live source before it counts as actionable.

Builds claim-to-evidence trace rows with statuses, gaps, tests, evals, reviews, and narrower non-claims.

Challenges the assumptions behind a change before an agent builds, merges, or releases, and names the one fact that would change the decision.

Attacks your own agent change, tool grant, dependency, model, or release on purpose, across risk types such as prompt injection, gaining extra power, unsafe output, and tool misuse, and records what you tried, what happened, and the leftover risk.

Routes a defect found in a shared or supplied artifact (a shared prompt, skill, dependency, model, eval, or template) to the downstream teams, agents, and releases that depend on it, not just a local fix.

Checks a critical agent action against its exact target, expected result, and stop condition before and after execution.

Records a ship, block, defer, or ship-with-risk decision that ties baseline, evidence status, residual risk, rollback, monitoring, and handoff together.

Checks that the way you cite source families, agencies, standards, or borrowed ideas is honest and does not claim too much.

Captures an agent run's tool calls, decision points, inputs, outputs, token use, and approval steps as a clear, repeatable record linked into the packet trace and verification record.

Checks dependencies, models, APIs, SaaS tools, generated files, and vendor claims against how you will actually use them, the proof you have, the gaps, and the release impact.

Hands off unfinished work with a closed-loop briefing of state, changed conditions, remaining scope, authority limits, and open evidence.

Verifies the delivered artifact a consumer actually receives (a figure, chart, PDF, SVG, screenshot, compiled build, or deployed response) by regenerating it fresh and directly observing the output, not the source that produces it, and ends in one honest verdict.

Decides which code, prompts, models, tools, dependencies, docs, tests, evals, releases, or claims need their approved state tracked.

PROVE Educate stage. Use after the verdict to lock in the approved baseline and turn operation into a lesson — record the baseline, OPEX/lessons, and any charter update into .nuclear/. Do not use to build product code, decide ship/block, or run the change.

PROVE Verdict stage. Use to make the ship / block / defer / ship-with-named-risk decision on the presented evidence — read-only and context-separated from the runner, but still coupled through the orchestrator and possibly model/resources. Do not use to build, gather new evidence, or write code.

PROVE Observe stage. Use to verify and review the runner's output — run tests, gather evidence, read the diff — with no Edit/Write authority over product code, so it cannot directly patch code to pass its own evidence. Do not use to build, plan, or decide ship/block.

PROVE Plan stage. Use to turn a request into a plan ready for human acceptance — question, discover, specify, plan — writing only to the change packet, never product code. Dispatch first, before any building. Do not use to edit code, run commands, or decide ship/block.

PROVE Run stage. Use AFTER a human explicitly authorizes execution of an accepted plan, to edit code, write files, and run commands strictly inside that plan's boundary. May fan out to parallel execution sub-agents on independent slices. Do not use to plan, to decide ship/block, or before the execut

Splits complex scope into a product-first hierarchy using the 100% rule to ensure deliverables are clearly defined without gaps or overlaps.

Prepares focused context for AI agents or reviewers by defining roles, goals, authority bounds, forbidden actions, and stop conditions.

Verifies factual assertions against primary sources and confirms them through a second independent method to ensure accuracy.

Reviews public documentation to ensure license, warranty, and compliance claims do not overreach the actual proven capabilities of a repository.

Records audited release decisions by evaluating evidence, residual risk, rollback plans, monitoring, and handoff requirements.

Verifies that citations of external sources, standards, or agencies are framed as influence rather than formal compliance or certification.

Identifies downstream ripple effects of controlled changes across documentation, tests, skills, and runtime systems to prevent stale artifacts.

Decides which code, prompts, models, tools, dependencies, docs, tests, evals, releases, or claims need their approved state tracked. Use when scoping what a change puts at risk or what must stay reviewable. Do not use for scratch work that nothing else depends on.

Brings abandoned or invalid change packets to a terminal state by completing, closing with a rationale, or deleting them.

Creates or updates Quick and Standard change records in Git to provide evidence-backed documentation for pull requests.

Determines whether an agent or a human holds authority for a specific action by evaluating reversibility, evidence, and consequences.

States what an agent intends to do and why before a critical or irreversible action, so a reviewer can challenge the thinking and not just the result. Use before deploys, migrations, public claims, or trust changes that deserve a stated intent, expected result, abort criteria, and backup. Do not use for routine reversible edits, and never treat the stated intent as proof the agent understood.

Verifies high-stakes agent actions by comparing intended targets and expected results against actual outcomes before claiming success.

Transfers responsibility between agents or owners using a closed-loop briefing of state, scope, authority limits, and remaining work.

Converts incidents, near misses, and errors into lasting control improvements by identifying root causes and updating safeguards.

Designs structured folder and file layouts based on work breakdown structures or disposition rules to ensure high cohesion and low coupling.

Builds traceable claim-to-evidence rows with statuses, gaps, and verification types to support release readiness decisions.

Challenges assumptions behind changes to prevent errors during builds, merges, or releases by identifying the single fact that would change a decision.

Selects an appropriate rigor mode for changes based on consequence, reversibility, and uncertainty to ensure appropriate oversight.

Records an agreed-upon version of a system, including its evidence, scope, accepted risks, and the triggers that would make the record obsolete.

Captures tool calls, decision points, inputs, outputs, and token usage to create a repeatable record for debugging, auditing, and cost review.

Routes defects found in shared artifacts to downstream consumers, teams, and releases to ensure widespread awareness and remediation.

Manages live production incidents using a stabilize-first approach with a named commander, fact-based timelines, and owned corrective actions.

Reviews code diffs or modules to prevent complexity drift by prioritizing deletion over rearrangement and enforcing architectural layering standards.

Convenes multiple independent reviewers on a frozen evidence packet to verify high-stakes verdicts through diverse axes of judgment and dissent preservation.

Prevents mission and standards drift by testing agent actions against a written mission anchor and forcing a decision to re-anchor, escalate, or stop.

Performs adversarial testing on agent changes to identify risks like prompt injection, tool misuse, and data leakage before release.

Maintains a standing register of known technical deficiencies to ensure they are owned, aged, and either fixed or formally risk-accepted.

Provides a structured routing and classification workflow for AI agents to perform preflight inspections and declare change modes before executing mutations.

Verifies produced artifacts like PDFs, charts, or UI screenshots by regenerating them fresh and observing the rendered output against a specification.

Evaluates dependencies, models, APIs, and SaaS tools by verifying critical characteristics against observed evidence rather than vendor claims.