stride

name	stride
description	Integration instructions for Stride kanban platform with AI-human collaboration workflows, client-side hooks, and task management API
license	MIT
compatibility	opencode
metadata	{"platform":"stride","api_base":"https://www.stridelikeaboss.com"}

STOP — orchestrator check

If you arrived here directly from a user prompt, you are in the wrong skill. Invoke stride:stride-workflow instead. Do not read further. Sub-skills are dispatched by the orchestrator only.

Orchestrator-First Pattern

For any Stride task work, the entry point is the stride-workflow skill. Every task should be claimed, explored, implemented, reviewed, and completed through that single lifecycle: claim → explore → implement → review → complete. The supporting skills (stride-claiming-tasks, stride-completing-tasks, stride-creating-tasks, stride-creating-goals, stride-enriching-tasks, stride-subagent-workflow) document API contracts and per-phase rules, but they are dispatched from inside stride-workflow, not invoked directly. This file is a legacy single-skill bundle of all Stride content; whatever runtime gating your host provides, the discipline of routing through stride-workflow is on the agent. When the user says "claim a task" or "complete this task", follow the lifecycle below — do not read or apply sub-skill contracts independently.

What I do

I provide comprehensive integration instructions for working with Stride, a kanban-based task management platform designed for AI-human collaboration. I cover:

• Client-side hook execution (before_doing, after_doing, before_review, after_review)
• Task claiming, completion, and creation workflows
• API endpoint usage and authentication
• Common mistakes to avoid
• Required task field structures
• Configuration file management

When to use me

Use this skill when you need to:

• Claim and complete tasks from Stride boards
• Create new tasks or goals with proper structure
• Execute workflow hooks at the right time
• Understand Stride's task schema and API
• Set up Stride configuration files (.stride.md, .stride_auth.md)

Hook Execution (MANDATORY)

Stride enforces workflow discipline through four client-side hooks that execute on your machine:

• before_doing (60s, blocking) - Execute before starting work (pull code, setup)
• after_doing (120s, blocking) - Execute after work, BEFORE calling complete endpoint (tests, lint)
• before_review (60s, blocking) - Execute BEFORE calling complete endpoint (create PR, docs)
• after_review (60s, blocking) - Execute after approval (merge, deploy)

CRITICAL: Execute BOTH after_doing AND before_review hooks BEFORE calling the task completion endpoint. The API requires both after_doing_result and before_review_result in the completion request. Hook validation failures must prevent task completion.

Top 5 Critical Mistakes to Avoid

1. Specifying Identifiers When Creating Tasks

DON'T:

{"identifier": "W47", "title": "Add dark mode", "type": "work"}

DO:

{"title": "Add dark mode", "type": "work"}

WHY: Identifiers are auto-generated (G1, W47, D12). Specifying them causes API rejection.

2. Using String Arrays for verification_steps

DON'T:

{"verification_steps": ["Run mix test", "Check coverage"]}

DO:

{
  "verification_steps": [
    {
      "step_type": "command",
      "step_text": "mix test path/to/test.exs",
      "expected_result": "All tests pass",
      "position": 0
    }
  ]
}

WHY: verification_steps must be array of objects with step_type, step_text, expected_result, and position.

3. Using Non-Array Values for testing_strategy Fields

DON'T:

{
  "testing_strategy": {
    "unit_tests": "Test the auth flow",
    "integration_tests": "Test end-to-end"
  }
}

DO:

{
  "testing_strategy": {
    "unit_tests": ["Test JWT generation", "Test token validation"],
    "integration_tests": ["Full auth flow with database"],
    "manual_tests": ["Verify login form works"],
    "edge_cases": ["Expired tokens", "Invalid credentials"],
    "coverage_target": "100% for auth module"
  }
}

WHY: unit_tests, integration_tests, and manual_tests must be arrays of strings.

4. Using Wrong Type Values

DON'T:

{"type": "task"}

DO:

{"type": "work"}

WHY: type must be exactly "work", "defect", or "goal" as strings.

5. Calling Complete Endpoint Before Executing BOTH Hooks

DON'T:

1. Finish work
2. Call PATCH /api/tasks/:id/complete
3. Execute hooks afterward

DO:

1. Finish work
2. Execute after_doing hook (tests, lint, build) — capture result
3. Execute before_review hook (create PR, docs) — capture result
4. Only if BOTH hooks succeed, call PATCH /api/tasks/:id/complete WITH both results

WHY: The API REQUIRES both after_doing_result and before_review_result parameters. Requests without them are rejected with 422 errors.

Essential Task Fields

Required:

• type - Exactly "work", "defect", or "goal"
• title - Clear, specific description

Critical:

• key_files - Prevents merge conflicts by marking which files are modified
• dependencies - Controls execution order using task identifiers or array indices
• verification_steps - Array of objects with proper structure
• testing_strategy - Must have arrays for unit/integration/manual tests

Strongly Recommended:

• description - WHY this matters and WHAT needs to be done
• complexity - "small", "medium", or "large"
• priority - "low", "medium", "high", or "critical"
• needs_review - Boolean controlling review requirement
• pitfalls - Array of what NOT to do
• patterns_to_follow - Code patterns to replicate (newline-separated string)
• acceptance_criteria - Definition of done (newline-separated string)
• why - Problem/value explanation
• what - Specific change description
• where_context - Location in code/UI

Code Patterns

Claiming a Task

# 1. Get next available task
curl -X GET https://www.stridelikeaboss.com/api/tasks/next \
  -H "Authorization: Bearer YOUR_TOKEN"

# 2. Execute before_doing hook FIRST (from .stride.md ## before_doing section)
# Capture exit_code, output, and duration_ms

# 3. Only if before_doing succeeds, claim the task WITH hook result
curl -X POST https://www.stridelikeaboss.com/api/tasks/claim \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "identifier": "W47",
    "agent_name": "OpenCode",
    "skills_version": "1.0",
    "before_doing_result": {
      "exit_code": 0,
      "output": "Already up to date.\nAll dependencies are up to date",
      "duration_ms": 1500
    }
  }'

# 4. Begin implementation work immediately

Completing a Task

# 1. Finish your work

# 2. Execute after_doing hook (from .stride.md ## after_doing section)
# Capture exit_code, output, duration_ms
# If hook fails, DO NOT proceed — fix issues and retry

# 3. Execute before_review hook (from .stride.md ## before_review section)
# Capture exit_code, output, duration_ms
# If hook fails, DO NOT proceed — fix issues and retry

# 4. Only if BOTH hooks succeed, mark complete WITH both results
curl -X PATCH https://www.stridelikeaboss.com/api/tasks/123/complete \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "agent_name": "OpenCode",
    "time_spent_minutes": 45,
    "completion_notes": "Implemented feature. All tests passing.",
    "completion_summary": "Added JWT auth with refresh tokens",
    "actual_complexity": "medium",
    "actual_files_changed": "lib/my_app/auth.ex, test/my_app/auth_test.exs",
    "skills_version": "1.0",
    "after_doing_result": {
      "exit_code": 0,
      "output": "Running tests...\n230 tests, 0 failures\nmix credo --strict\nNo issues found",
      "duration_ms": 45678
    },
    "before_review_result": {
      "exit_code": 0,
      "output": "Creating pull request...\nPR #123 created",
      "duration_ms": 2340
    }
  }'

Complete endpoint required fields:

• agent_name (string) - Name of the completing agent
• time_spent_minutes (integer) - Actual time spent
• completion_notes (string) - Summary of what was done
• completion_summary (string) - Brief summary for tracking
• actual_complexity (enum) - "small", "medium", or "large"
• actual_files_changed (string) - Comma-separated file paths (NOT an array)
• after_doing_result (object) - Hook result with exit_code, output, duration_ms
• before_review_result (object) - Hook result with exit_code, output, duration_ms

Creating a Task

{
  "type": "work",
  "title": "[Verb] [What] [Where/Context]",
  "description": "Clear explanation of WHY this matters and WHAT needs to be done",
  "complexity": "medium",
  "priority": "high",
  "needs_review": true,
  "key_files": [
    {
      "file_path": "lib/my_app/auth.ex",
      "note": "Add JWT validation",
      "position": 0
    }
  ],
  "dependencies": ["W45", "W46"],
  "verification_steps": [
    {
      "step_type": "command",
      "step_text": "mix test test/auth_test.exs",
      "expected_result": "All tests pass",
      "position": 0
    }
  ],
  "testing_strategy": {
    "unit_tests": ["Test JWT generation", "Test token validation"],
    "integration_tests": ["Full auth flow with database"],
    "manual_tests": ["Verify login form works"],
    "edge_cases": ["Expired tokens", "Invalid credentials"],
    "coverage_target": "100% for auth module"
  },
  "acceptance_criteria": "Users can log in with email/password\nJWT tokens are generated correctly\nExpired tokens are rejected",
  "patterns_to_follow": "Follow existing auth patterns in lib/my_app/auth/*.ex",
  "pitfalls": ["Don't hardcode secrets", "Don't skip token validation"],
  "why": "Users need secure authentication to access protected resources",
  "what": "Implement JWT-based authentication system",
  "where_context": "lib/my_app/auth/ directory and related controllers"
}

Creating a Goal with Nested Tasks

{
  "type": "goal",
  "title": "Implement User Authentication System",
  "description": "Complete authentication system with JWT, password reset, and 2FA",
  "complexity": "large",
  "tasks": [
    {
      "type": "work",
      "title": "Create user schema and migration",
      "complexity": "small",
      "key_files": [{"file_path": "priv/repo/migrations/xxx_create_users.exs", "note": "User table", "position": 0}],
      "dependencies": []
    },
    {
      "type": "work",
      "title": "Implement JWT authentication",
      "complexity": "medium",
      "key_files": [{"file_path": "lib/my_app/auth.ex", "note": "JWT logic", "position": 0}],
      "dependencies": [0]
    },
    {
      "type": "work",
      "title": "Add password reset functionality",
      "complexity": "medium",
      "dependencies": [1]
    }
  ]
}

Stride Configuration Files

Your project should have these files:

1. .stride_auth.md (NEVER commit - add to .gitignore)

   • Contains API URL and bearer token
   • Template available from onboarding endpoint

2. .stride.md (commit to version control)

   • Contains hook execution scripts
   • Four sections: before_doing, after_doing, before_review, after_review

Documentation Links

• Onboarding: https://www.stridelikeaboss.com/api/agent/onboarding
• Task Writing Guide: https://raw.githubusercontent.com/cheezy/kanban/refs/heads/main/docs/TASK-WRITING-GUIDE.md
• API Reference: https://raw.githubusercontent.com/cheezy/kanban/refs/heads/main/docs/api/README.md
• Hook Execution: https://raw.githubusercontent.com/cheezy/kanban/refs/heads/main/docs/AGENT-HOOK-EXECUTION-GUIDE.md
• AI Workflow: https://raw.githubusercontent.com/cheezy/kanban/refs/heads/main/docs/AI-WORKFLOW.md

Quick Reference

Workflow: before_doing hook → claim WITH result → work → after_doing hook → before_review hook → complete WITH both results → [if needs_review=false: after_review hook → claim next, else: stop]

API Base: https://www.stridelikeaboss.com

Authentication: Authorization: Bearer YOUR_TOKEN

Never commit: .stride_auth.md (contains secrets)

Always commit: .stride.md (contains hooks)