Agent skill

Nav Workflow

by qf-studio in qf-studio/navigator

Unified workflow orchestration for substantial tasks. An agent skill from qf-studio/navigator.

MITAuto-check: notesAgent Workflows

Install Nav Workflow

skills CLI
$ npx skills add qf-studio/navigator --skill nav-workflow -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install qf-studio/navigator nav-workflow --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/qf-studio/navigator.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/nav-workflow .claude/skills/nav-workflow && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
nav-workflow
GitHub stars
355
Token cost
~2.7k tokens
SKILL.md length
762 words
Files
7
Skills in repo
32
Repo updated
First seen
Licence
MIT

At a glance

Unified workflow orchestration for substantial tasks. An agent skill from qf-studio/navigator.

  • Works in 5 steps: Analyze Request → Check Skill Match → Initialize Task Mode → …
  • Tasks that involve Agent instruction files
  • SKILL.md covers Why This Exists, How It Works, When to Invoke and Configuration, plus 8 more sections
  • Runs Python scripts from its folder; calls python3

What it does

Nav Workflow is an agent skill from qf-studio/navigator. Unified workflow orchestration for substantial tasks. Auto-detects complexity, defers to matching skills, or provides phase-based execution. Solves workflow conflicts between skills, loop mode, and CLAUDE.md.

Its SKILL.md is about 2.7k tokens, which your agent loads only when the skill is triggered. The skill folder holds 7 other files (for example `functions/complexity_detector.py`, `functions/phase_indicator.py` and `functions/skill_detector.py`).

It sits in Agent Workflows, covering Agent instruction files. The repository describes itself as: Finish What You Start — Context engineering for Claude Code. Sessions last 20+ exchanges instead of crashing at 7. The licence is MIT.

When your agent uses it

  • Tasks that involve Agent instruction files

Example prompts

  • “/nav-workflow”

Requirements

  • Python 3
  • Pre-approved tools (allowed-tools): Read, Write, Edit, Bash, Grep, Glob, AskUserQuestion

Workflow steps

5 steps, taken from the step headings in SKILL.md.

  1. Analyze Request
  2. Check Skill Match
  3. Initialize Task Mode
  4. Execute Phases
  5. Complete Task Mode

What it can do on your machine

Read from SKILL.md and the folder at commit 3bb9eac. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Read
    • Write
    • Edit
    • Bash
    • Grep
    • Glob
    • AskUserQuestion

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships script files (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Nav Workflow loads about 2.7k tokens when it runs. Until then it costs about 55 tokens; SKILL.md has 762 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~55
When it runs · the whole SKILL.md, loaded when a task matches
~2.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
    allowed-tools: Read, Write, Edit, Bash, Grep, Glob, AskUserQuestion

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from qf-studio/navigator at commit 3bb9eac, republished under its MIT licence (© qf-studio). 762 words, ~2,740 tokens.

Download SKILL.mdSave it as .claude/skills/nav-workflow/SKILL.md (or your agent's skills folder). This skill also uses 6 other files; get the full folder from GitHub.
name
nav-workflow
description
Unified workflow orchestration for substantial tasks. Auto-detects complexity, defers to matching skills, or provides phase-based execution. Solves workflow conflicts between skills, loop mode, and CLAUDE.md.
allowed-tools
Read, Write, Edit, Bash, Grep, Glob, AskUserQuestion
version
1.0.0

Navigator Task Mode Skill

Unified workflow orchestration that coordinates between skills, loop mode, and direct execution based on task complexity and type.

Why This Exists

Navigator had three disconnected workflow systems:

  1. Skills (frontend-component, etc.) - have mini-workflows (Step 1 → Step 7)
  2. Loop Mode - separate phase system (INIT → COMPLETE)
  3. CLAUDE.md - documents workflow nobody enforces

Result: Conflicts when multiple systems try to run.

Solution: Task Mode acts as a coordinator - detecting when skills should handle workflow vs when to provide standalone phase guidance.

How It Works

User Request
    ↓
TASK MODE (this skill)
    ├─ Simple task? → Direct execution (no overhead)
    ├─ Skill matches? → Let skill run (it has workflow)
    └─ Substantial, no skill? → Task Mode phases

When to Invoke

Auto-invoke when:

  • User starts substantial work (3+ steps expected)
  • No obvious skill match (not "create component", "add endpoint", etc.)
  • Request involves planning, refactoring, or multi-file changes
  • Loop mode is disabled but structured execution needed

DO NOT invoke if:

  • Trivial task (typo fix, single line change)
  • Skill will clearly handle it (component creation, endpoint, migration, etc.)
  • User says "quick", "just do", "simple fix"
  • Already in Task Mode or Loop Mode

Configuration

Task Mode settings in .agent/.nav-config.json:

json
{
  "task_mode": {
    "enabled": true,
    "auto_detect": true,
    "defer_to_skills": true,
    "complexity_threshold": 0.5,
    "show_phase_indicator": true
  }
}

Options:

  • enabled: Master switch for Task Mode
  • auto_detect: Auto-detect complexity (vs explicit invocation only)
  • defer_to_skills: Let matching skills handle their own workflow
  • complexity_threshold: Score (0-1) required to activate
  • show_phase_indicator: Show phase banners during execution

Execution Steps

Step 1: Analyze Request

Run complexity detection:

bash
python3 functions/complexity_detector.py \
  --request "{USER_REQUEST}" \
  --context "{RECENT_CONTEXT}"

Complexity signals:

  • Multi-file changes expected (+0.3)
  • Planning language ("implement", "refactor", "add feature") (+0.2)
  • Vague requirements needing research (+0.2)
  • Cross-system changes (frontend+backend) (+0.3)
  • Testing requirements mentioned (+0.1)

Simple task signals:

  • Single file mentioned (-0.3)
  • Fix/typo/update language (-0.2)
  • Specific location given (-0.2)
  • "Quick" or "simple" mentioned (-0.3)

Decision:

IF complexity_score < threshold:
  → Direct execution (exit Task Mode)
Step 2: Check Skill Match

Run skill detection:

bash
python3 functions/skill_detector.py \
  --request "{USER_REQUEST}" \
  --available-skills "{SKILLS_LIST}"

Skill matching rules:

  • "create component" → frontend-component
  • "add endpoint" → backend-endpoint
  • "database migration" → database-migration
  • "write test" → backend-test/frontend-test
  • etc.

Decision:

IF matching_skill AND defer_to_skills:
  → Show: "Detected: {SKILL_NAME} skill will handle this"
  → Exit Task Mode (skill has workflow)
Step 3: Initialize Task Mode

If substantial task, no skill match:

Display activation:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
TASK MODE ACTIVATED
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Task: {TASK_SUMMARY}
Complexity: {SCORE} (threshold: {THRESHOLD})
Skills matched: None (Task Mode will orchestrate)

Phases:
  ○ RESEARCH - Understand requirements
  ○ PLAN - Create implementation strategy
  ○ IMPL - Execute changes
  ○ VERIFY - Test and validate
  ○ COMPLETE - Commit and document

Starting RESEARCH phase...
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Step 4: Execute Phases

RESEARCH Phase:

  • Use Task agent for codebase exploration
  • State the Ideal Final Result (the no-new-code version) and the reuse inventory (what already does ≥80% of this)
  • If the brief declared a Contradiction, query prior resolutions (graph_manager.py --action contradictions --filter "<terms>") and run the nav-triz divergence step: three candidates from different separation modes, one recommended, before PLAN
  • Identify affected files
  • Find existing patterns
  • Document unknowns

Show phase transition:

bash
python3 functions/phase_indicator.py \
  --phase "RESEARCH" \
  --status "complete" \
  --next "PLAN"
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PHASE: RESEARCH → PLAN
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Research completed:
  ✓ Found {N} related files
  ✓ Identified patterns in {LOCATION}
  ✓ Dependencies mapped

Moving to PLAN phase...
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

PLAN Phase:

  • Create TodoWrite items
  • Identify order of changes
  • Document approach

IMPL Phase:

  • Execute planned changes
  • Follow project patterns
  • Write tests as appropriate

VERIFY Phase:

  • Run tests
  • Type check
  • Build validation
  • Run nav-simplify (if enabled)

COMPLETE Phase:

  • Commit changes
  • Update documentation
  • Close tickets (if PM configured)
  • Suggest compact
Step 5: Complete Task Mode

Display completion:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
TASK MODE COMPLETE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Task: {TASK_SUMMARY}
Phases:
  ✓ RESEARCH - {DURATION}
  ✓ PLAN - {DURATION}
  ✓ IMPL - {DURATION}
  ✓ VERIFY - {DURATION}
  ✓ COMPLETE

Summary:
- {FILES_CHANGED} files changed
- {TESTS_ADDED} tests added
- Committed: {COMMIT_SHA}

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Predefined Functions

functions/complexity_detector.py

Analyzes request to determine complexity score.

Usage:

bash
python3 functions/complexity_detector.py \
  --request "Refactor the auth system to use JWT" \
  --context "Working on user management"

Returns:

json
{
  "complexity_score": 0.7,
  "signals": {
    "multi_file": true,
    "planning_language": true,
    "cross_system": false
  },
  "recommendation": "task_mode",
  "reason": "Refactoring task with multi-file scope"
}
functions/skill_detector.py

Checks if a skill should handle this request.

Usage:

bash
python3 functions/skill_detector.py \
  --request "Add a login component" \
  --available-skills '["frontend-component", "backend-endpoint", "database-migration"]'

Returns:

json
{
  "matching_skill": "frontend-component",
  "confidence": 0.95,
  "triggers": ["create component", "add component"],
  "defer": true,
  "reason": "Request matches frontend-component skill triggers"
}
functions/phase_indicator.py

Generates phase transition displays.

Usage:

bash
python3 functions/phase_indicator.py \
  --phase "IMPL" \
  --status "in_progress" \
  --progress 60 \
  --details '{"files_changed": 3, "tests_written": 2}'

Returns: Formatted phase indicator block.


Integration with Navigator

Show full SKILL.md (309 more words)Show less
With Loop Mode

Task Mode is lighter weight than Loop Mode:

  • Loop Mode: Strict iteration control, stagnation detection, EXIT_SIGNAL
  • Task Mode: Phase guidance, skill coordination, no strict gates

When to use which:

  • Loop Mode: "Run until done", autonomous iteration
  • Task Mode: Substantial task, need structure but not strict iteration

Can coexist: Loop Mode wraps Task Mode phases if both active.

With Skills

Task Mode defers to skills by default:

  • Skill has its own workflow (Steps 1-7)
  • Task Mode doesn't add overhead
  • Just shows "Skill X will handle this"

Override: Set defer_to_skills: false to always use Task Mode phases.

With nav-simplify

Simplification runs during VERIFY phase:

  • After tests pass
  • Before committing
  • Respects simplification.enabled config
With Autonomous Completion

COMPLETE phase triggers autonomous protocol:

  • Commit changes
  • Archive documentation
  • Close tickets
  • Create markers

Comparison Table

AspectDirect ExecutionTask ModeLoop Mode
ComplexityLowMedium-HighHigh
Phase trackingNoneVisual phasesStrict phases
Iteration controlNoneNoneEXIT_SIGNAL
Skill coordinationNoneDefersIndependent
Best forQuick fixesFeaturesAutonomous work

Examples

Example 1: Simple Fix (Direct Execution)
User: "Fix the typo in README"

→ Complexity: 0.1 (below threshold)
→ Direct execution (no Task Mode overhead)
Example 2: Component Creation (Skill Defers)
User: "Create a UserProfile component"

→ Complexity: 0.6
→ Skill match: frontend-component (0.95 confidence)
→ Task Mode defers: "frontend-component skill will handle this"
→ Skill executes its own Step 1-7 workflow
Example 3: Refactoring (Task Mode Active)
User: "Refactor auth to use JWT instead of sessions"

→ Complexity: 0.8
→ Skill match: None
→ Task Mode activates
→ RESEARCH: Explore current auth implementation
→ PLAN: Document JWT migration steps
→ IMPL: Execute changes across files
→ VERIFY: Run tests, simplify code
→ COMPLETE: Commit, document
Example 4: With Loop Mode
User: "Run until done: implement user roles"

→ Loop Mode activated (explicit trigger)
→ Task Mode phases guide each iteration:
    Iteration 1: RESEARCH phase
    Iteration 2: PLAN + IMPL phases
    Iteration 3: IMPL + VERIFY phases
    Iteration 4: COMPLETE + EXIT_SIGNAL

Error Handling

Config not found:

Task Mode config not found in .nav-config.json.
Using defaults: auto_detect=true, threshold=0.5

Skill detection fails:

  • Fall back to Task Mode (conservative)
  • Log warning but continue

Phase detection ambiguous:

  • Default to current or next logical phase
  • Show reasoning for user

Success Criteria

Task Mode succeeds when:

  • Simple tasks execute without overhead
  • Skills handle their matched requests
  • Substantial tasks get phase structure
  • No conflicts between systems
  • User sees clear progress indicators

Verification Checklist

After implementation:

  • "Add login component" → frontend-component runs (not Task Mode)
  • "Refactor auth system" → Task Mode runs (no matching skill)
  • "Fix typo" → Direct execution (not substantial)
  • Phase indicators show during Task Mode
  • Skills continue working unchanged
  • Loop Mode can wrap Task Mode phases

This skill provides unified workflow orchestration, resolving conflicts between Navigator's multiple workflow systems.

© qf-studio, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 6 other files in skills/nav-workflow of qf-studio/navigator.

  • SKILL.md
  • functions/complexity_detector.py
  • functions/phase_indicator.py
  • functions/skill_detector.py
  • functions/test_complexity_detector.py
  • functions/test_phase_indicator.py
  • functions/test_skill_detector.py

Open the folder on GitHubat commit 3bb9eac

Compare with similar skills

Nav Workflow next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Nav Workflow compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Nav Workflow this skillqf-studio/navigator355—~2.7kAutomated safety check: NotesMIT
Using Agent Skillsaddyosmani/agent-skills104k4 repos~2.4kAutomated safety check: PassMIT
Claude ReflectBayramAnnakov/claude-reflect1.8k2 repos~627Automated safety check: PassMIT
Neat-Freak Knowledge CloseoutKKKKhazix/khazix-skills21k—~1.9kAutomated safety check: PassMIT
Writing For Agentsbestofjs/bestofjs3.1k18 repos~2.7kAutomated safety check: PassMIT
Task Observerrebelytics/one-skill-to-rule-them-all3.2k1 repos~11kAutomated safety check: PassCC-BY-4.0

Similar skills

  • Using Agent Skills

    addyosmani/agent-skills

    Meta-skill for choosing which workflow skill fits the task at hand, plus always-on habits: surface assumptions, stop on confusion, push back, keep it simple and stay in scope.

    104k GitHub starsUsed in 4 repos~2.4k tokens
    Agent WorkflowsAuto-check passed
  • Claude Reflect

    BayramAnnakov/claude-reflect

    Self-learning system that captures corrections during sessions and reminds users to run /reflect to update CLAUDE.md.

    1.8k GitHub starsUsed in 2 repos~627 tokens
    Agent WorkflowsAuto-check passed
  • Neat-Freak Knowledge Closeout

    KKKKhazix/khazix-skills

    Brings project docs, agent rule files, authorized memory and leftover workspace files back in line with what the code and runtime actually do at the end of a work session.

    21k GitHub stars~1.9k tokensUpdated 9 days ago
    Agent WorkflowsAuto-check passed
  • Writing For Agents

    bestofjs/bestofjs

    Writing documents for agents. An agent skill from bestofjs/bestofjs.

    3.1k GitHub starsUsed in 18 repos~2.7k tokens
    Agent WorkflowsAuto-check passed
  • Task Observer

    rebelytics/one-skill-to-rule-them-all

    Monitors task execution for skill improvement opportunities.

    3.2k GitHub starsUsed in 1 repo~11k tokens
    Agent WorkflowsAuto-check passed
  • SkillOpt Sleep Cycle

    microsoft/SkillOpt

    Official

    Runs an on-demand or nightly sleep cycle that reviews past Claude Code sessions and proposes validated updates to CLAUDE.md and skills.

    18k GitHub stars~2.3k tokensUpdated 4 days ago
    Agent WorkflowsAuto-check passed

More from qf-studio/navigator

All 32 skills in this repo
  • Nav Start

    qf-studio/navigator

    Load Navigator documentation navigator when starting development session, resuming work, or beginning new feature.

    355 GitHub stars~4.7k tokensUpdated 2 days ago
    Auto-check: notes
  • Backend Endpoint

    qf-studio/navigator

    Create REST/GraphQL API endpoint with validation, error handling, and tests.

    355 GitHub stars~4.5k tokensUpdated 2 days ago
    Auto-check: notes
  • Backend Test

    qf-studio/navigator

    Generate backend tests (unit, integration, mocks) for existing code.

    355 GitHub stars~1.5k tokensUpdated 2 days ago
    Auto-check: notes
  • Database Migration

    qf-studio/navigator

    Create database migration with schema changes and rollback. An agent skill from qf-studio/navigator.

    355 GitHub stars~3.7k tokensUpdated 2 days ago
    Auto-check: notes
  • Frontend Component

    qf-studio/navigator

    Create React/Vue component with TypeScript, tests, and styles.

    355 GitHub stars~4.5k tokensUpdated 2 days ago
    Auto-check: notes
  • Frontend Test

    qf-studio/navigator

    Generate frontend component tests (React Testing Library, Vue Test Utils, snapshot) for existing components.

    355 GitHub stars~1.6k tokensUpdated 2 days ago
    Auto-check: notes

Categories

Questions about Nav Workflow

What does Nav Workflow do?

Unified workflow orchestration for substantial tasks. An agent skill from qf-studio/navigator. Nav Workflow is an agent skill from qf-studio/navigator. Unified workflow orchestration for substantial tasks.

When should I use Nav Workflow?

Nav Workflow fits situations like: tasks that involve Agent instruction files.

How do I install Nav Workflow in Claude Code?

Run `npx skills add qf-studio/navigator --skill nav-workflow -a claude-code`. Or copy the skill folder (skills/nav-workflow in qf-studio/navigator) into .claude/skills/nav-workflow in your project. Claude Code loads it when a task matches its description.

How do I install Nav Workflow in Codex?

Run `npx skills add qf-studio/navigator --skill nav-workflow -a codex`. Or copy the skill folder (skills/nav-workflow in qf-studio/navigator) into .agents/skills/nav-workflow in your project. Codex loads it when a task matches its description.

Can I use Nav Workflow in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add qf-studio/navigator --skill nav-workflow -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/nav-workflow, .gemini/skills/nav-workflow, .github/skills/nav-workflow and .opencode/skills/nav-workflow in your project.

What does Nav Workflow need to run?

Going by SKILL.md and its folder, Nav Workflow needs Python for the scripts in its folder and the command-line tools its instructions call (python3). Our summary lists: Python 3. Its frontmatter pre-approves these tools: Read, Write, Edit, Bash, Grep, Glob, AskUserQuestion.

Does Nav Workflow access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Nav Workflow safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Nav Workflow use?

Nav Workflow is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Nav Workflow use?

About 2.7k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Nav Workflow?

Skills that share tags, products or a category with Nav Workflow: Using Agent Skills (addyosmani/agent-skills, 104k stars), Claude Reflect (BayramAnnakov/claude-reflect, 1.8k stars), Neat-Freak Knowledge Closeout (KKKKhazix/khazix-skills, 21k stars) and Writing For Agents (bestofjs/bestofjs, 3.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Nav Workflow?

qf-studio (a GitHub organization) maintains it in qf-studio/navigator, which has 355 GitHub stars. The repository holds 32 skills in this directory. The repository was last updated on October 8, 2026.

Source: qf-studio/navigator on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.