Agent skill
task
Structured implementation process for features, chores, refactors, and docs. Use when implementing non-bugfix work.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/task-dohernandez-claude-project-skill
SKILL.md
Task
Purpose
Structured implementation methodology for non-bugfix work (features, chores, refactors, docs). Ensures work is properly scoped, planned, implemented following project patterns, and verified before completion.
Quick Reference
- Phases: Scope -> Plan -> Implement -> Verify
- Key Rule: Understand before building
- Output: Working implementation with changes reviewed
Task Phases
+-----------------------------------------------------------------------------+
| TASK WORKFLOW |
+-----------------------------------------------------------------------------+
| 1. SCOPE -> 2. PLAN -> 3. IMPLEMENT -> 4. VERIFY |
+-----------------------------------------------------------------------------+
Phase T1: Scope
Goal: Understand what needs to be done and define boundaries.
Procedure:
- Read the ticket/request description carefully
- Identify the core requirement (what must be delivered)
- Identify out-of-scope items (what NOT to do)
- Clarify any ambiguities with user
- Document the scope
Output: Clear understanding of deliverables
## Scope
- **Goal:** Add retry logic to the E2E runner script
- **Deliverables:**
- Configurable retry count in run-e2e.sh
- Proper exit code propagation
- **Out of scope:**
- Changes to the GitHub Actions workflow triggers
- Dashboard or reporting changes
Key Rule: Don't start planning until scope is clear. Ask questions early.
Phase T2: Plan
Goal: Design the approach before writing code.
Procedure:
- Explore the relevant parts of the codebase
- Identify files to create/modify
- Consider how changes interact with existing scripts and workflows
- Document the plan
- Get user approval on plan before implementing
Output: Implementation plan with files and approach
## Plan
- **Area:** E2E runner orchestration
- **Files to modify:**
- scripts/run-e2e.sh (add retry loop)
- .github/workflows/run-e2e.yml (pass retry count input)
- **Approach:** Wrap test execution in retry loop with configurable count
Key Rule: Get user approval on plan before implementing.
Phase T3: Implement
Goal: Write code following project patterns.
Procedure:
- Follow existing code patterns and conventions in the project
- Work in small increments
- Keep changes minimal and focused on the scope
- Use proper shell scripting practices (set -euo pipefail, quoting, etc.)
- Follow YAML best practices for workflow files
Guidelines:
- Match existing code style and patterns
- Don't over-engineer
- Keep shell scripts defensive (error handling, input validation)
- Use consistent naming with existing files
Output: Working code changes
Key Rule: Minimal changes only. Don't refactor unrelated code.
Phase T4: Verify
Goal: Ensure implementation is complete and correct.
Procedure:
- Review the diff of all changes against the original scope
- Run shellcheck on modified shell scripts (if available)
- Validate YAML syntax on modified workflow files
- Check for common issues: unquoted variables, missing error handling, hardcoded values
- Verify no unintended side effects in related files
Output: Changes reviewed, no obvious errors, ready for commit
# Verification checks (as applicable)
shellcheck scripts/run-e2e.sh # Lint shell scripts
yamllint .github/workflows/*.yml # Validate YAML syntax
git diff --stat # Review scope of changes
Key Rule: Don't skip the review step. Always check the diff against the original scope.
Flow Diagram
+-------------------+
| Request received |
+---------+---------+
|
v
+-------------------+ +-------------------+
| T1: SCOPE |---->| Unclear? |
| Understand req | | Ask user |
+---------+---------+ +-------------------+
| Clear
v
+-------------------+ +-------------------+
| T2: PLAN |---->| Need approval |
| Design approach | | Show plan |
+---------+---------+ +-------------------+
| Approved
v
+-------------------+ +-------------------+
| T3: IMPLEMENT |---->| Issues? |
| Write code | | Iterate |
| |<----| |
+---------+---------+ +-------------------+
| Complete
v
+-------------------+ +-------------------+
| T4: VERIFY |---->| Problems found? |
| Review changes | | Fix and re-check |
| |<----| |
+---------+---------+ +-------------------+
| All clear
v
+-------------------+
| Ready for |
| commit |
+-------------------+
Task Context
Track progress in workflow context:
{
"type": "feat",
"task": {
"phase": "implement",
"scopeConfirmed": true,
"planApproved": true,
"filesModified": ["scripts/run-e2e.sh"]
}
}
Anti-Patterns
| Don't | Do Instead |
|---|---|
| Start coding immediately | Scope and plan first |
| Plan in isolation | Get user approval on plan |
| Change unrelated code | Minimal changes only |
| Skip reviewing the diff | Always review changes against scope |
| Ignore lint/syntax errors | Fix all errors before commit |
Automation
See skill.yaml for patterns and procedures.
See sharp-edges.yaml for common implementation pitfalls.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?