ReddRadar
Agent skill
phase-gate
Manages pre-phase and post-phase agent selection gates. Core skill for agent-first pipeline execution.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/phase-gate
SKILL.md
Phase Gate Skill
Activation
Triggered automatically at every phase transition:
[ACTIVATE:PHASE_GATE_V1]
Parameters:
gate_type: pre | post
phase: 1-12
Purpose
Facilitate human-Claude conversations for agent selection at phase boundaries:
- Pre-Phase Gate: Select executor agents before phase begins
- Post-Phase Gate: Select auditor agents after phase completes
Agent-First Philosophy
In the agent-first pipeline:
- Agents execute phases, not skills
- Skills are playbooks that guide agent behavior
- Humans approve agent selection at every boundary
- PhD-experts can be invented when gaps exist
Execution Flow
Pre-Phase Gate
╔═══════════════════════════════════════════════════════════════════════╗
║ PRE-PHASE GATE: Phase {N} - {Phase Name} ║
╠═══════════════════════════════════════════════════════════════════════╣
║ ║
║ Phase {N} is about to begin. ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ DEFAULT EXECUTORS │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ PRIMARY: │ ║
║ │ • {executor-1} - {description} │ ║
║ │ │ ║
║ │ SUPPORT: │ ║
║ │ • {executor-2} - {description} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ OPTIONAL SPECIALISTS │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ • {optional-1} - {description} │ ║
║ │ • {optional-2} - {description} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ TECH-SPECIFIC (from PRD) │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ • {tech-agent-1} - for {technology} │ ║
║ │ • {tech-agent-2} - for {technology} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
╠═══════════════════════════════════════════════════════════════════════╣
║ [A] Accept defaults ║
║ [C] Customize selection (add/remove agents) ║
║ [I] Invent new PhD-expert for this phase ║
║ [V] View agent details ║
╚═══════════════════════════════════════════════════════════════════════╝
Pre-Phase Gate Steps
-
Load Phase Configuration
javascriptconst phaseConfig = require('.claude/phase-agents.json').phases[phaseNum]; const techStack = require('.claude/prd-context.json').techStack; -
Determine Default Executors
- Read
phaseConfig.executors.default - Add tech-specific agents based on PRD
- Present to user
- Read
-
Handle User Choice
- Accept: Proceed with defaults
- Customize: Interactive add/remove
- Invent: Activate
AGENT_INVENTOR_V1 - View: Show agent system prompts
-
Lock Executor Selection
json{ "phase": 6, "gate": "pre", "executors": ["spec-generator", "typescript-pro", "reactjs-expert"], "locked_at": "2025-01-15T10:30:00Z", "approved_by": "human" } -
Emit Signal
[SIGNAL:PRE_PHASE_{N}_COMPLETE] [ACTIVATE:{PHASE_SKILL}]
Post-Phase Gate
╔═══════════════════════════════════════════════════════════════════════╗
║ POST-PHASE GATE: Phase {N} - {Phase Name} ║
╠═══════════════════════════════════════════════════════════════════════╣
║ ║
║ Phase {N} has completed. ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ PHASE OUTPUTS │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ • {output-1}: {description} │ ║
║ │ • {output-2}: {description} │ ║
║ │ • {output-3}: {description} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ DEFAULT AUDITORS │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ • {auditor-1} - {what they verify} │ ║
║ │ • {auditor-2} - {what they verify} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
║ ┌────────────────────────────────────────────────────────────────┐ ║
║ │ RECOMMENDED BASED ON OUTPUTS │ ║
║ ├────────────────────────────────────────────────────────────────┤ ║
║ │ • {recommended-1} - because {reason} │ ║
║ └────────────────────────────────────────────────────────────────┘ ║
║ ║
╠═══════════════════════════════════════════════════════════════════════╣
║ [A] Accept defaults ║
║ [C] Customize auditor selection ║
║ [I] Invent specialist auditor for this output ║
║ [S] Skip audit (not recommended) ║
╚═══════════════════════════════════════════════════════════════════════╝
Post-Phase Gate Steps
-
Analyze Phase Outputs
- Read phase output files
- Determine what was produced
- Identify potential risk areas
-
Recommend Auditors
- Start with
phaseConfig.auditors.default - Add recommendations based on output content:
- Security-sensitive →
security-auditor - Data operations →
data-integrity-auditor - API changes →
api-compatibility-auditor - Performance-critical →
performance-auditor
- Security-sensitive →
- Start with
-
Handle User Choice
- Accept: Run default auditors
- Customize: Add/remove auditors
- Invent: Create specialist auditor
- Skip: Require confirmation, log risk
-
Lock Auditor Selection
json{ "phase": 6, "gate": "post", "auditors": ["spec-reviewer", "security-auditor"], "locked_at": "2025-01-15T14:30:00Z", "approved_by": "human" } -
Execute Audit
- Run each auditor against phase outputs
- Collect findings
- Determine pass/fail
-
Handle Audit Results
If PASS:
╔═══════════════════════════════════════════════════════════════════════╗ ║ AUDIT PASSED ✓ ║ ╠═══════════════════════════════════════════════════════════════════════╣ ║ ║ ║ Phase {N} outputs verified by: ║ ║ ✓ {auditor-1}: PASSED ║ ║ ✓ {auditor-2}: PASSED ║ ║ ║ ║ Proceeding to Phase {N+1} ║ ║ ║ ╚═══════════════════════════════════════════════════════════════════════╝If FAIL:
╔═══════════════════════════════════════════════════════════════════════╗ ║ AUDIT FAILED ✗ ║ ╠═══════════════════════════════════════════════════════════════════════╣ ║ ║ ║ Phase {N} audit found issues: ║ ║ ║ ║ {auditor-1}: FAILED ║ ║ • Issue 1: {description} ║ ║ • Issue 2: {description} ║ ║ ║ ║ {auditor-2}: PASSED ║ ║ ║ ╠═══════════════════════════════════════════════════════════════════════╣ ║ [R] Remediate issues and re-audit ║ ║ [O] Override (requires justification) ║ ║ [B] Back to phase execution ║ ╚═══════════════════════════════════════════════════════════════════════╝ -
Emit Signal
[SIGNAL:POST_PHASE_{N}_COMPLETE] [SIGNAL:PHASE_{N}_AUDITED] [ACTIVATE:PHASE_GATE_V1] gate_type=pre, phase={N+1}
State Management
Track gate states in .claude/phase-state.json:
{
"current_phase": 6,
"gates": {
"1": {
"pre": { "status": "complete", "executors": [...], "timestamp": "..." },
"post": { "status": "complete", "auditors": [...], "passed": true }
},
"6": {
"pre": { "status": "complete", "executors": [...] },
"post": { "status": "pending" }
}
},
"invented_agents": [
{ "name": "graphql-phd-expert", "phase": 6, "role": "executor" }
]
}
Quick Confirm Mode
For phases where defaults are well-suited:
Phase 5 Pre-Gate: [taskmaster-integrator, assignment-agent]
Quick confirm? [Y/n]:
User can press Enter to accept, or 'n' to enter full selection mode.
Agent Availability Check
Before presenting agents, verify availability (agents are bundled in agents/):
# Check if agent exists in local agents directory
if [ ! -f "agents/*/{agent}.md" ]; then
echo "Agent not found in local pool"
# Consider inventing a PhD-expert if needed
fi
# Sync to .claude/agents/ if needed
./scripts/pull-agents.sh --config .claude/agent-config.json
Emergency Overrides
For critical situations:
[OVERRIDE:SKIP_AUDIT]
Justification: {required}
Risk acknowledged: {yes|no}
Logged at: .claude/audit-log.json
Overrides are logged and flagged for review.
Signals
| Signal | Meaning |
|---|---|
PRE_PHASE_{N}_COMPLETE |
Executors selected, phase can begin |
POST_PHASE_{N}_COMPLETE |
Auditors selected, audit can begin |
PHASE_{N}_AUDITED |
Audit complete (pass or overridden) |
AGENT_SELECTION_TIMEOUT |
No response, defaults applied |
Integration with Phase Skills
Phase skills now receive executor context:
{
"phase": 6,
"skill": "spec-gen",
"executors": [
{"agent": "spec-generator", "role": "primary"},
{"agent": "typescript-pro", "role": "tech-specific"}
],
"mode": "agent-first"
}
The skill becomes the "playbook" that agents follow, not the executor.
Completion
Outputs:
- Updated
.claude/phase-state.json - Agent selection logged
- Audit results recorded (for post-gate)
Transitions to:
- Phase execution (after pre-gate)
- Next phase pre-gate (after post-gate audit passes)
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
majiayu000/claude-skill-registry
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
majiayu000/claude-skill-registry
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
majiayu000/claude-skill-registry
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
majiayu000/claude-skill-registry
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
majiayu000/claude-skill-registry
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
majiayu000/claude-skill-registry
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?