Task Reference
Nightshift includes 59 built-in tasks organized into 6 categories. Each task has a cost tier, risk level, and default cooldown interval.
Use nightshift task list to browse tasks, or nightshift task show <name> to see details for a specific task.
PR Tasks — "It's done — here's the PR"
Fully formed, review-ready artifacts. These tasks create branches and open pull requests.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
lint-fix | Linter Fixes | Automatically fix linting errors and style issues | Low | Low | 24h |
bug-finder | Bug Finder & Fixer | Identify and fix potential bugs in code | High | Medium | 72h |
auto-dry | Auto DRY Refactoring | Identify and refactor duplicate code | High | Medium | 7d |
skill-groom | Skill Grooming | Audit and update project-local agent skills to match the current codebase | High | Medium | 7d |
api-contract-verify | API Contract Verification | Verify API contracts match implementation | Medium | Low | 7d |
backward-compat | Backward-Compatibility Checks | Check and ensure backward compatibility | Medium | Low | 7d |
build-optimize | Build Time Optimization | Optimize build configuration for faster builds | High | Medium | 7d |
docs-backfill | Documentation Backfiller | Generate missing documentation | Low | Low | 7d |
commit-normalize | Commit Message Normalizer | Standardize commit message format | Low | Low | 24h |
changelog-synth | Changelog Synthesizer | Generate changelog from commits | Low | Low | 7d |
release-notes | Release Note Drafter | Draft release notes from changes | Low | Low | 7d |
adr-draft | ADR Drafter | Draft Architecture Decision Records | Medium | Low | 7d |
td-review | TD Review Session | Review open td reviews, fix obvious bugs, create tasks for bigger issues | High | Medium | 72h |
note
td-review is disabled by default and must be explicitly opted in via tasks.enabled. It requires the td integration to be enabled (see Integrations).
Analysis Tasks — "Here's what I found"
Completed analysis with conclusions. These tasks produce reports without modifying code.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
doc-drift | Doc Drift Detector | Detect documentation that's out of sync with code | Medium | Low | 72h |
semantic-diff | Semantic Diff Explainer | Explain the semantic meaning of code changes | Medium | Low | 72h |
dead-code | Dead Code Detector | Find unused code that can be removed | Medium | Low | 72h |
dependency-risk | Dependency Risk Scanner | Analyze dependencies for security and maintenance risks | Medium | Low | 72h |
test-gap | Test Gap Finder | Identify areas lacking test coverage | Medium | Low | 72h |
test-flakiness | Test Flakiness Analyzer | Identify and analyze flaky tests | Medium | Low | 72h |
logging-audit | Logging Quality Auditor | Audit logging for completeness and quality | Medium | Low | 72h |
metrics-coverage | Metrics Coverage Analyzer | Analyze metrics instrumentation coverage | Medium | Low | 72h |
perf-regression | Performance Regression Spotter | Identify potential performance regressions | Medium | Low | 72h |
cost-attribution | Cost Attribution Estimator | Estimate resource costs by component | Medium | Low | 72h |
security-footgun | Security Foot-Gun Finder | Find common security anti-patterns | Medium | Low | 72h |
pii-scanner | PII Exposure Scanner | Scan for potential PII exposure | Medium | Low | 72h |
privacy-policy | Privacy Policy Consistency Checker | Check code against privacy policy claims | Medium | Low | 72h |
schema-evolution | Schema Evolution Advisor | Analyze database schema changes | Medium | Low | 72h |
event-taxonomy | Event Taxonomy Normalizer | Normalize event naming and structure | Medium | Low | 72h |
roadmap-entropy | Roadmap Entropy Detector | Detect roadmap scope creep and drift | Medium | Low | 72h |
bus-factor | Bus-Factor Analyzer | Analyze code ownership concentration | Medium | Low | 72h |
knowledge-silo | Knowledge Silo Detector | Identify knowledge silos in the team | Medium | Low | 72h |
Options Tasks — "Here are options"
These tasks surface judgment calls, tradeoffs, and design forks for human review.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
task-groomer | Task Groomer | Refine and clarify task definitions | Medium | Low | 7d |
guide-improver | Guide/Skill Improver | Suggest improvements to guides and skills | Medium | Low | 7d |
idea-generator | Idea Generator | Generate improvement ideas for the codebase | Medium | Low | 7d |
tech-debt-classify | Tech-Debt Classifier | Classify and prioritize technical debt | Medium | Low | 7d |
why-annotator | Why Does This Exist Annotator | Document the purpose of unclear code | Medium | Low | 7d |
edge-case-enum | Edge-Case Enumerator | Enumerate potential edge cases | Medium | Low | 7d |
error-msg-improve | Error-Message Improver | Suggest better error messages | Medium | Low | 7d |
slo-suggester | SLO/SLA Candidate Suggester | Suggest SLO/SLA candidates | Medium | Low | 7d |
ux-copy-sharpener | UX Copy Sharpener | Improve user-facing text | Medium | Low | 7d |
a11y-lint | Accessibility Linting | Non-checkbox accessibility analysis | Medium | Low | 7d |
service-advisor | Should This Be a Service Advisor | Analyze service boundary decisions | High | Medium | 7d |
ownership-boundary | Ownership Boundary Suggester | Suggest code ownership boundaries | Medium | Low | 7d |
oncall-estimator | Oncall Load Estimator | Estimate oncall load from code changes | Medium | Low | 7d |
Safe Tasks — "I tried it safely"
Required execution or simulation but left no lasting side effects.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
migration-rehearsal | Migration Rehearsal Runner | Rehearse migrations without side effects | Very High | High | 14d |
contract-fuzzer | Integration Contract Fuzzer | Fuzz test integration contracts | Very High | High | 14d |
golden-path | Golden-Path Recorder | Record golden path test scenarios | High | Medium | 14d |
perf-profile | Performance Profiling Runs | Run performance profiling | High | Medium | 14d |
allocation-profile | Allocation/Hot-Path Profiling | Profile memory allocation and hot paths | High | Medium | 14d |
Map Tasks — "Here's the map"
Pure context laid out cleanly. These tasks document and visualize system structure.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
visibility-instrument | Visibility Instrumentor | Instrument code for observability | High | Medium | 7d |
repo-topology | Repo Topology Visualizer | Visualize repository structure | Medium | Low | 7d |
permissions-mapper | Permissions/Auth Surface Mapper | Map permissions and auth surfaces | Medium | Low | 7d |
data-lifecycle | Data Lifecycle Tracer | Trace data lifecycle through the system | Medium | Low | 7d |
feature-flag-monitor | Feature Flag Lifecycle Monitor | Monitor feature flag usage and lifecycle | Medium | Low | 7d |
ci-signal-noise | CI Signal-to-Noise Scorer | Score CI signal vs noise ratio | Medium | Low | 7d |
historical-context | Historical Context Summarizer | Summarize historical context of code | Medium | Low | 7d |
Emergency Tasks — "For when things go sideways"
Artifacts you hope to never need. These tasks prepare for incident response.
| Task | Name | Description | Cost | Risk | Cooldown |
|---|---|---|---|---|---|
runbook-gen | Runbook Generator | Generate operational runbooks | High | Medium | 30d |
rollback-plan | Rollback Plan Generator | Generate rollback plans for changes | High | Medium | 30d |
postmortem-gen | Incident Postmortem Draft Generator | Draft incident postmortem documents | High | Medium | 30d |
Cost Tiers
| Tier | Token Range | Task Count |
|---|---|---|
| Low | 10–50k tokens | 5 |
| Medium | 50–150k tokens | 39 |
| High | 150–500k tokens | 13 |
| Very High | 500k+ tokens | 2 |
Risk Levels
| Level | Description | Task Count |
|---|---|---|
| Low | Safe to run autonomously | 44 |
| Medium | May modify code; review recommended | 13 |
| High | Significant execution; careful review required | 2 |