plan-review

name	plan-review
description	Multi-persona review gate for implementation plans. Validates architectural soundness, scope boundaries, and coding standards compliance before the harvest skill decomposes a plan into work items.
argument-hint	[path to plan file in docs/exec-plans/]

Plan Review Gate

Validates implementation plans through multi-persona review before the harvest skill decomposes them into work items. This gate prevents flawed plans from generating flawed work hierarchies.

Subagent Depth Constraint

This skill spawns reviewer subagents. Those subagents are leaf executors and MUST NOT spawn their own subagents. Maximum depth: plan-review skill → persona subagent (1 hop).

Severity Scale

Level	Meaning	Gate action
P0	Plan will produce unshippable or unsafe code (missing security, broken contracts, impossible scope)	Block harvest
P1	Plan has a high-impact gap that will cause significant rework (missing requirements, wrong decomposition, absent verification)	Block harvest
P2	Plan has a moderate gap (edge case coverage, missing test scenario, suboptimal decomposition)	Record as backlog follow-up
P3	Plan has a minor improvement opportunity (wording, optional optimization)	Advisory

Use the same severity conventions as the code review skill adapted for plan artifacts rather than code diffs.

Agent-Intercom Communication

When the agent-intercom capability pack is installed, call ping at session start. If reachable, broadcast at every step. If unreachable, warn the operator that visibility is degraded.

Event	Level	Message prefix
Review start	info	[PLAN-REVIEW] Starting review of: {plan_path}
Persona spawned	info	[SPAWN] {persona_name} for plan review
Persona returned	info	[RETURN] {persona_name}: {finding_count} findings
Merge complete	info	[PLAN-REVIEW] Merged: {total_findings} findings ({p0} P0, {p1} P1, {p2} P2, {p3} P3)
Gate decision	success/error	[PLAN-REVIEW] Gate: {PASS|FAIL|ADVISORY}
Review appended	success	[PLAN-REVIEW] Review appended to: {plan_path}

Inputs

• plan_path: (Required) Path to the plan file (docs/exec-plans/{YYYY-MM-DD}-{slug}-plan.md).

If no path is provided, search docs/exec-plans/ for the most recent *-plan.md file and confirm with the operator.

Output

Review findings are appended to the plan file as a ## Plan Review section, not written as a separate file. The plan-review skill produces a gate decision (PASS, ADVISORY, or FAIL) that is recorded in the appended section. When the compact-context skill later consolidates the plan, it merges the plan and appended reviews into a decided-plan.

Reviewer Personas

Spawn all always-on personas and any triggered cross-model personas. Use different models when available to force genuine diversity of critique.

Always-On Personas (same model as caller)

Persona Subagent	Focus
Constitution Reviewer	Map plan units against constitutional principles. Flag violations.
Rust Reviewer	Evaluate proposed type signatures, error handling patterns, package boundaries, and verification steps.
Scope Boundary Auditor	Verify units stay within declared scope. Detect scope creep, YAGNI, unnecessary complexity.
Learnings Researcher	Search docs/compound/ for prior solutions relevant to the plan's scope. Report P0 if the plan contradicts a known past resolution; P1 if it ignores a highly relevant prior solution.

Cross-Model Personas (different model when available)

Persona Subagent	Focus	Suggested Model
Architecture Strategist	Cohesion, coupling, module boundaries, dependency chains.	Different from caller
Agent-Native Parity Reviewer	Plans that expose MCP tools, agent-facing actions, or user/agent parity-sensitive workflows.	Different from caller
Security Lens Reviewer (security-lens-reviewer.agent.md)	Plans that touch auth/authz systems, API surfaces, sensitive data stores, external integrations, or secrets management.	Different from caller

If cross-model invocation is not available, run all personas with the caller's model. Multi-model is preferred but not blocking.

Workflow

Step 1: Load and Parse Plan

1. Read the plan file from docs/exec-plans/.
2. Extract implementation units, dependency graph, decisions, risks, hardening signals, and whether a ## Plan Hardening section is present.
3. When strict-safety is enabled and the plan contains a ## Plan Hardening section, also extract any ProposedAction / ActionRisk entries.
4. If the plan references an origin document, read that too for context.
5. Broadcast: [PLAN-REVIEW] Starting review of: {plan_path}

Step 2: Spawn Reviewer Subagents

Spawn all always-on personas plus the cross-model personas whose trigger conditions are met. Each receives:

• The full plan content
• The origin requirements doc (if any)
• The project's coding standards and conventions (reference .github/instructions/constitution.instructions.md)
• Instructions to return structured findings

Trigger conditions for cross-model personas:

• Architecture Strategist: always triggered
• Agent-Native Parity Reviewer: triggered when the plan exposes MCP tools, agent-facing actions, or user/agent parity-sensitive workflows
• Security Lens Reviewer: triggered when the plan touches authentication or authorization systems, API surfaces, sensitive data stores, external integrations crossing trust boundaries, or secrets and credentials management

Broadcast each spawn.

Step 3: Collect and Merge Findings

As each persona returns:

1. Broadcast the return with finding count
2. Collect all findings into a unified list
3. Deduplicate: merge findings that identify the same issue from different perspectives
4. Assign final severity (use the more conservative severity when personas disagree)

Step 4: Gate Decision

Condition	Decision	Action
Plan shows hardening signals but lacks plan hardening or equivalent high-risk detail	FAIL	Return the plan to plan-harden or manual revision before harvest.
Strict-safety enabled, hardening present, but risky actions lack ProposedAction / ActionRisk classification	FAIL	Plans with hardening signals must classify risky actions explicitly when strict-safety is active.
Any P0 or P1 findings	FAIL	Present findings to user. Plan must be revised before proceeding to harvest.
P2 findings only	ADVISORY	Present findings to user. User decides: revise or proceed.
P3 findings only or none	PASS	Log findings as advisory. Proceed to harvest.

Broadcast the gate decision.

Step 5: Append Review to Plan

Append a ## Plan Review section to the plan file with:

• Gate decision and rationale
• Whether plan hardening was required and whether that requirement was satisfied
• All findings organized by severity
• Specific recommendations for addressing P0/P1 issues
• Acknowledgment of P2/P3 items for awareness
• Runtime verification and operational closure gaps called out explicitly when missing

The review is appended (not written as a separate file) so that the plan and its review travel together as a single artifact. The compact-context skill later consolidates this into a decided-plan.

Quality Criteria

• Every implementation unit is reviewed by at least the always-on personas
• The gate decision correctly reflects finding severities
• Findings include actionable recommendations
• The review is appended to the plan before the gate decision is communicated
• Plans with hardening signals are failed when hardening is missing or materially incomplete
• Plans that touch runtime surfaces are checked for verification and closure readiness

Model Routing

This skill operates at Tier 2 (Standard) — plan review coordination and finding assembly.

Generated by autoharness | Template: plan-review/SKILL.md.tmpl