ln-810-performance-optimizer
Paths: File paths (
references/,../ln-*) are relative to this skill directory.
Type: L2 Domain Coordinator Category: 8XX Optimization
Performance Optimizer
Runtime-backed multi-cycle optimization coordinator. Profiles, researches, validates, and executes optimization hypotheses until target reached, plateau detected, or budget exhausted.
Inputs
| Input | Required | Description |
|---|---|---|
target |
Yes | endpoint, function, or pipeline to optimize |
observed_metric |
Yes | current performance problem |
target_metric |
No | user or research-derived target |
max_cycles |
No | default 3 |
Purpose & Scope
- Detect whether optimization is the right tool
- Run iterative cycles:
profile -> gate -> research -> target -> context -> validate -> execute - Preserve optimization artifacts under
.hex-skills/optimization/{slug}/and runtime state under.hex-skills/optimization/runtime/runs/{run_id}/ - Resume deterministically from the last checkpointed phase
- Keep cycle summaries machine-readable
Runtime Contract
MANDATORY READ: Load references/ci_tool_detection.md
MANDATORY READ: Load references/coordinator_runtime_contract.md, references/optimization_runtime_contract.md, references/coordinator_summary_contract.md
Runtime CLI:
node references/scripts/optimization-runtime/cli.mjs start --slug {slug} --manifest-file .hex-skills/optimization/{slug}/manifest.json
node references/scripts/optimization-runtime/cli.mjs status --slug {slug}
node references/scripts/optimization-runtime/cli.mjs record-worker-result --payload '{...}'
node references/scripts/optimization-runtime/cli.mjs record-summary --payload '{...}'
node references/scripts/optimization-runtime/cli.mjs record-cycle --payload '{...}'
node references/scripts/optimization-runtime/cli.mjs checkpoint --phase PHASE_8_EXECUTE --payload '{...}'
node references/scripts/optimization-runtime/cli.mjs advance --to PHASE_9_CYCLE_BOUNDARY
Runtime Layout
Runtime state is run-scoped, while optimization artifacts stay slug-scoped:
.hex-skills/optimization/
runtime/active/ln-810/{slug}.json
runtime/runs/{run_id}/manifest.json
runtime/runs/{run_id}/state.json
runtime/runs/{run_id}/checkpoints.json
runtime/runs/{run_id}/history.jsonl
runtime-artifacts/runs/{run_id}/optimization-coordinator/ln-810--{slug}.json
runtime-artifacts/runs/{run_id}/optimization-worker/{worker}--{child_identifier}.json
{slug}/context.md
{slug}/ln-814-log.tsv
Workflow
Phase 0: Preflight
- Validate:
- target identifiable
- observed metric provided
- git clean state
- test infrastructure exists
- Detect stack and optional service topology.
- Derive slug.
- Build manifest with:
slugtargetobserved_metrictarget_metricexecution_modecycle_config
- Start runtime and checkpoint
PHASE_0_PREFLIGHT.
Phase 1: Parse Input
- Normalize the problem statement.
- Set or defer
target_metric. - Checkpoint
PHASE_1_PARSE_INPUT.
Phase 2: Profile
- Compute deterministic child metadata:
identifier=ln-811--{slug}--cycle-{current_cycle}- child
run_id - exact
summaryArtifactPath=.hex-skills/runtime-artifacts/runs/{parent_run_id}/optimization-worker/ln-811--{slug}--cycle-{current_cycle}.json
- Checkpoint
PHASE_2_PROFILEwithchild_run. - Invoke
ln-811-performance-profilerwith the childrunIdand exactsummaryArtifactPath. - Read the emitted
optimization-workersummary envelope from the exact artifact path. - Record the worker summary with
record-worker-result.
Phase 3: Wrong Tool Gate
Evaluate profiler output:
| Gate | Meaning | Action |
|---|---|---|
PROCEED |
optimization work is justified | continue |
CONCERNS |
measurements usable but imperfect | continue with warning |
BLOCK |
wrong tool, already optimized, or infrastructure-bound | aggregate and exit |
WAIVED |
user overrides BLOCK |
continue with explicit waiver |
Rules:
- cycle 1
BLOCK-> finish as diagnostic result - cycle 2+
BLOCKdue toalready_optimizedorwithin_industry_norm-> finish as successful stop
Checkpoint PHASE_3_WRONG_TOOL_GATE with:
gate_verdictstop_reasonwhen blockedfinal_resultwhen terminal
Phase 4: Research
- Compute deterministic child metadata for
ln-812. - Checkpoint
PHASE_4_RESEARCHwithchild_run. - Invoke
ln-812-optimization-researcherwith the childrunIdand exactsummaryArtifactPath. - Read and record the emitted
optimization-workersummary envelope. - If no hypotheses remain, stop after aggregate/report.
Phase 5: Set Target
- Resolve target metric:
- user-specified target wins
- otherwise use research target with confidence
- otherwise default to 50% improvement
- Checkpoint
PHASE_5_SET_TARGETwithtarget_metric.
Phase 6: Write Context
- Build
.hex-skills/optimization/{slug}/context.md. - Include:
- problem statement
- performance map
- target metrics
- hypotheses and conflicts
- local codebase findings
- previous cycles
- Checkpoint
PHASE_6_WRITE_CONTEXTwithcontext_file.
Phase 7: Validate Plan
- Compute deterministic child metadata for
ln-813. - Checkpoint
PHASE_7_VALIDATE_PLANwithvalidation_verdictandchild_run. - Invoke
ln-813-optimization-plan-validatorwith the childrunIdand exactsummaryArtifactPath. - Read and record the emitted
evaluation-coordinatorsummary envelope. - If verdict is
NO_GO, pause runtime until user resolves or waives.
Phase 8: Execute
execution_mode=execute:
- Compute deterministic child metadata for
ln-814. - Checkpoint
PHASE_8_EXECUTEwithchild_run. - Invoke
ln-814-optimization-executorwith the childrunIdand exactsummaryArtifactPath. - Read and record the emitted
optimization-workersummary envelope.
execution_mode=plan_only:
- Do not run
ln-814. - Checkpoint
PHASE_8_EXECUTEasskipped_by_mode.
Phase 9: Cycle Boundary
- Record the cycle summary with
record-cycle. - Evaluate stop conditions:
- target met
- plateau
- max cycles reached
- no new hypotheses
- If continuing:
- merge previous branch when needed
- increment
current_cycle - checkpoint
PHASE_9_CYCLE_BOUNDARY - advance back to
PHASE_2_PROFILE
- If stopping:
- checkpoint
PHASE_9_CYCLE_BOUNDARYwithstop_reason - advance to
PHASE_10_AGGREGATE
- checkpoint
Phase 10: Aggregate
- Aggregate all cycle summaries from runtime state.
- Compute cumulative improvement.
- Checkpoint
PHASE_10_AGGREGATE.
Phase 11: Report
- Produce final report with:
- per-cycle summary
- cumulative improvement
- final result
- gap analysis when target not met
- Checkpoint
PHASE_11_REPORTwith:report_ready=truefinal_result
- Record the
optimization-coordinatorsummary envelope withrecord-summary. - Complete runtime only after the report checkpoint and coordinator summary exist.
Worker Invocation (MANDATORY)
| Phase | Skill | Purpose |
|---|---|---|
| 2 | ln-811-performance-profiler |
Build measured performance map |
| 4 | ln-812-optimization-researcher |
Research hypotheses and targets |
| 7 | ln-813-optimization-plan-validator |
Validate feasibility via evaluation-platform review (L2 Coordinator) |
| 8 | ln-814-optimization-executor |
Execute optimization strike and bisect |
Skill(skill: "ln-811-performance-profiler")
Skill(skill: "ln-812-optimization-researcher")
Agent(... Skill(skill: "ln-813-optimization-plan-validator"))
Agent(... Skill(skill: "ln-814-optimization-executor"))
TodoWrite format (mandatory)
- Start ln-810 runtime (pending)
- Run profiler and record summary (pending)
- Apply Wrong Tool Gate (pending)
- Run researcher and record summary (pending)
- Set target metric (pending)
- Write optimization context (pending)
- Validate plan and record summary (pending)
- Execute or skip by mode (pending)
- Record cycle boundary (pending)
- Aggregate results and write final report (pending)
Critical Rules
- Runtime state is the optimization orchestration SSOT.
- Worker outputs are consumed only through summary JSON artifacts.
plan_onlyis a first-class execution mode, not an informal branch.NO_GOfromln-813must pause runtime until explicitly resolved.- Cycle history lives in runtime state, not in chat memory.
- A terminal diagnostic result is still a valid
DONEorchestration outcome.
Definition of Done
- Runtime started and preflight/input checkpoints recorded
- Profiler and researcher summaries recorded deterministically
- Wrong Tool Gate result checkpointed
- Target metric and context file checkpointed
- Validator summary recorded;
NO_GOhandled viaPAUSEDwhen needed - Executor summary recorded or
skipped_by_modecheckpointed - Cycle boundary recorded for every completed cycle
- Aggregate and final report checkpoints recorded
- Runtime completed with final result and resume-free terminal state
Phase 12: Meta-Analysis
MANDATORY READ: Load references/meta_analysis_protocol.md
Skill type: optimization-coordinator. Run after phases complete. Output to chat using the optimization-coordinator format.
Reference Files
references/coordinator_runtime_contract.mdreferences/optimization_runtime_contract.mdreferences/coordinator_summary_contract.md../ln-811-performance-profiler/SKILL.md../ln-812-optimization-researcher/SKILL.md../ln-813-optimization-plan-validator/SKILL.md../ln-814-optimization-executor/SKILL.md
Version: 3.0.0 Last Updated: 2026-03-15
More from levnikolaevich/claude-code-skills
ln-624-code-quality-auditor
Checks cyclomatic complexity, nesting, long methods, god classes, O(n2), N+1 queries, constants management. Use when auditing code quality.
413ln-620-codebase-auditor
Use when auditing the codebase through the evaluation platform with mandatory research, coordinated domain audit workers, and structured summaries.
357ln-626-dead-code-auditor
Checks unreachable code, unused imports/variables/functions, commented-out code, unsupported patterns. Use when auditing dead code.
354ln-621-security-auditor
Checks hardcoded secrets, SQL injection, XSS, insecure deps, input validation. Use when auditing security.
337ln-782-test-runner
Executes all test suites and reports results with coverage. Use when verifying that test infrastructure works after bootstrap.
330ln-634-test-coverage-auditor
Identifies missing tests for critical paths (money, security, data integrity, core flows). Use when auditing test coverage gaps.
324