task-classification
Skill8 starsChanged 17 days agoMITView on GitHub
This skill should be used when classifying tasks, routing to agents, determining complexity, or handling task routing, agent selection, verification requirements, and multi-component changes.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- ZadAnthony/meta-workflowtheir own version, first seen Oct 11, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- ZadAnthony/meta-workflow★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- agent-behavior-constraintsSkill
- agentsview-usageSkill
- evidence-driven-workSkill
- explainer-design-systemSkill
- exploration-strategySkill
- graphify-usageSkill
- personal-kb-usageSkill
- prompt-refinementSkill
- skill-agent-mappingSkill
- verification-gatesSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 29
Err on the side of caution: When uncertain, classify higher rather than lower
- Distinctive: line 67
Read-only investigation focused on understanding code or gathering information.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Task Classification
Overview
Task classification is the foundational skill that enables intelligent routing of user requests to the appropriate agent(s) within the multi-agent orchestration system. This skill analyzes incoming tasks to determine:
- Complexity Level: How many files, components, or systems are affected
- Risk Assessment: Potential for introducing bugs, regressions, or breaking changes
- Agent Assignment: Which specialized agent(s) should handle the task
- Verification Requirements: Whether automated testing and review are needed
When to Use This Skill
Apply task classification when:
- A new user request arrives that requires action
- The scope of work is unclear or potentially complex
- Multiple approaches could solve the same problem
- Determining if verification gates are required
- Delegating work to specialized agents
Key Principles
- Err on the side of caution: When uncertain, classify higher rather than lower
- Consider hidden complexity: Database migrations, API changes, and authentication have amplified risk
- Account for dependencies: Changes that affect shared code require broader verification
- Respect the verification mandate: Implementation and Complex tasks always require verification
Task Categories
Tasks are classified into five primary categories:
| Category | Files | Risk | Primary Agent | Verification |
| Trivial | 0-1 | Low | Direct | None |
| Exploratory | N/A | Low | Riko | None |
| Implementation | 2-5 | Medium | Loid | Lawliet + Alphonse |
| Complex | 5+ | High | Full orchestration | Lawliet + Alphonse (+ Codex) |
| Research | N/A | Low | Riko + WebSearch | None |
Inside
/orchestrate, the tier feeds the orchestrator-chosen Execution Profile. Implementation/complex/trivial map to fast / standard / thorough, where Lawliet and Alphonse always run and Codex is skipped under fast.research/exploratory(or an explicit report request) take the report-only route: Riko, plus Senku unless plain exploratory, then the Research Short-Circuit with no Phases 3–5. Seecommands/orchestrate.md→ Execution Profile and research-short-circuit.md.
Parallel Work Eligibility
Agent Teams and /team-orchestrate are deprecated. Inside /orchestrate, parallelism is handled by parallel background dispatch — the orchestrator decides it itself (e.g., Phase 4 review and Phase 5 verification always run in parallel). For decomposable tasks, split work across parallel Loid dispatches only when:
- Task independence (no shared files or dependencies)
- File ownership clarity (exclusive write access per subtask)
- Cost-benefit (time savings justify coordination overhead)
Trivial Tasks
Simple, low-risk tasks handled directly without agent delegation.
Examples: Answering code questions, single-line fixes, documentation comments
Indicators: "what", "why", "how" questions; localized changes; no logic changes
Exploratory Tasks
Read-only investigation focused on understanding code or gathering information.
Examples: Finding usages, tracing data flow, understanding architecture
Agent: Riko (Explorer) with Grep, Glob, Read tools
Implementation Tasks
Standard development tasks involving code changes across a moderate number of files.
Examples: New API endpoints, feature flags, bug fixes, refactoring functions
Agents: Loid (Executor) -> Lawliet (Reviewer) + Alphonse (Verifier) in parallel
Complex Tasks
High-impact changes affecting multiple components, requiring full orchestration.
Examples: Major refactoring, security changes, database migrations, cross-service features
Agents: Riko -> Senku -> Loid -> Lawliet + Alphonse (+ Codex), review and verification in parallel
Research Tasks
Information-gathering requiring external research or documentation lookup.
Examples: Best practices research, library evaluation, documentation lookup
Agent: Riko (Explorer) with WebSearch
Agent Quick Reference
| Agent | Role | Model | Key Tools |
| Riko | Explorer | Sonnet | Read, Grep, Glob, Bash*, WebSearch, WebFetch |
| Senku | Planner | Opus | Read, Grep, Glob, Write (plan/report files only) |
| Loid | Executor | Sonnet | Read, Write, Edit, Grep, Glob, Bash |
| Lawliet | Reviewer | Sonnet | Read, Grep, Glob, Bash |
| Alphonse | Verifier | Sonnet | Bash, Read, Grep |
| Speedwagon | Authoring | Sonnet | Read, Grep, Glob, Write†, Edit†, Bash‡ — dispatch for /explain authoring |
- Riko's Bash access is limited to AST analysis tools only (ast-grep, tree-sitter, language parsers)
For detailed agent routing, see references/agent-selection-matrix.md.
Quick Classification Process
Three-Question Decision
- Is this read-only? -> Yes: Exploratory (Riko)
- How many files? -> 0-1: Trivial, 2-5: Implementation, 5+: Complex
- High-risk domain? -> Yes: Complex (regardless of file count)
Override Conditions
These conditions force Complex classification:
- Security-sensitive changes (authentication, authorization, encryption)
- Database migrations or schema changes
- Changes to payment or billing systems
- Multi-service coordination required
- Breaking API changes
For detailed classification steps, see references/classification-process.md.
Risk Amplifiers
| Domain | Risk Increase | Reason |
| Authentication | Critical | Security breach potential |
| Database schema | Critical | Data loss, migration complexity |
| API contracts | High | Breaking changes for consumers |
| Shared utilities | Medium | Wide blast radius |
| Payment/billing | Critical | Financial impact |
| Configuration | Medium | Runtime behavior changes |
Verification Requirements
| Task Type | Type Check | Lint | Unit Tests | Integration | Review |
| Trivial | - | - | - | - | - |
| Exploratory | - | - | - | - | - |
| Implementation | Required | Required | Required | Optional | Required (Lawliet) |
| Complex | Required | Required | Required | Required | Required |
| Research | - | - | - | - | - |
In the /orchestrate parallel flow, type check and lint are run by Lawliet (Alphonse reports them COVERED (Lawliet)); Alphonse runs tests and build.
Keyword Indicators
| Keywords | Likely Classification |
| "what", "why", "how", "explain" | Trivial or Exploratory |
| "find", "search", "where", "trace" | Exploratory |
| "fix", "update", "add", "change" | Implementation |
| "refactor", "redesign", "migrate" | Complex |
| "research", "best practices", "compare" | Research |
| "all", "every", "entire", "across" | Complex (scope indicator) |
Additional Resources
Reference Files
- references/agent-selection-matrix.md - Detailed agent routing tables
- references/classification-process.md - Step-by-step classification
- references/decision-flowchart.md - Visual decision guides
- references/classification-best-practices.md - Best practices and pitfalls
- references/classification-heuristics.md - Edge case handling
- references/deep-dive-synthesis.md - Synthesis for /deep-dive command
- references/research-short-circuit.md -
/orchestrateresearch/exploratory short-circuit and plan-approved continuation
Examples
- examples/classification-examples.md - Worked classification examples
Related Skills
- agent-behavior-constraints: Model routing, tool access, and behavioral guardrails
- verification-gates: Verification requirements and pass criteria
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.