/implementing-tasks
Implementing a single planned task (Task N.M.T) end-to-end: selects the right backend agent by language and service type, drives one TDD RED->GREEN turn, then verifies coverage, lint, license headers, runtime, and delivery before handoff. Runs as Gate 0 before
$ npx -y skills add LerianStudio/ring --skill implementing-tasks --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/implementing-tasks
Context preview
The summary Claude sees to decide when to auto-load this skill.
Implementing a single planned task (Task N.M.T) end-to-end: selects the right backend agent by language and service type, drives one TDD RED->GREEN turn, then verifies coverage, lint, license headers, runtime, and delivery before handoff. Runs as Gate 0 before
SKILL.md
implementing-tasks.SKILL.mdname: ring:implementing-tasks
description: "Implementing a single planned task (Task N.M.T) end-to-end: selects the right backend agent by language and service type, drives one TDD RED->GREEN turn, then verifies coverage, lint, license headers, runtime, and delivery before handoff. Runs as Gate 0 before ring:reviewing-code. Use to drive ONE task inside an already-running cycle. Skip when asked to implement a whole plan.md or multiple tasks (use ring:running-dev-cycle)."
Code Implementation (Gate 0)
When to use
- Gate 0 of development cycle
- Tasks loaded at initialization
- Ready to write code
Skip when
- Not inside a development cycle (ring:running-dev-cycle or ring:planning-backend-refactor)
- Task is documentation-only, configuration-only, or non-code
- Implementation already completed for the current gate
Sequence
**Runs before:** ring:reviewing-code
Related
**Complementary:** ring:running-dev-cycle, ring:test-driven-development, ring:reviewing-code
You orchestrate. Agents implement. Select the agent, prepare the prompt, track state, validate outputs.
Step 1: Validate Input
The unit of work is a single task (Task N.M.T, e.g. Task 1.1.1) from the phased plan. Its requirements arrive inline from the orchestrator, which reads the task block under its epic in plan.md: Context, Implementation vision, Files, Verification, and Done when.
Required: `unit_id` (the task's `Task N.M.T` id), `requirements` (the task block), `language` (go|typescript|python), `service_type` (api|worker|batch|cli|frontend|bff). Optional: `technical_design`, `existing_patterns`, `project_rules_path` (default: `docs/PROJECT_RULES.md`).
STOP if any required input is missing.
Step 2: Validate Prerequisites
Check `PROJECT_RULES.md` exists at `project_rules_path` → STOP if not found.
**Agent selection:**
| Language | Service Type | Agent | |----------|--------------|-------| | go | api, worker, batch, cli | ring:backend-go | | typescript | api, worker | ring:backend-ts | | typescript | frontend, bff | ring:bff-ts |
Step 3: Gate 0 — TDD (RED → GREEN)
Dispatch the selected agent ONCE. The agent writes the failing test, captures the RED failure output as evidence, then implements to GREEN — in one turn.
Task:
subagent_type: "{selected_agent}"
description: "Gate 0 TDD (RED→GREEN) for {unit_id}"
prompt: |
## TDD: write a failing test, then make it pass
unit_id: {unit_id}
requirements: {requirements}
language: {language}
service_type: {service_type}
Standards: load via state.cached_standards or WebFetch Ring standards for the language.
Project rules: {project_rules_path}
## Frontend TDD policy (React/Next.js only)
Visual-only components (layout, styling, animations): RED not required — report
"Visual-only → RED skipped; visual checks apply in the frontend flow."
Behavioral components (hooks, validation, state, conditional rendering, API): RED required.
## Multi-Tenant (Go only)
Implement DUAL-MODE from the start. Use resolvers for all resources
(tmcore.GetPGContext, tmcore.GetMBContext, etc.) — they work transparently
in single-tenant and multi-tenant mode. Load multi-tenant.md for patterns.
## Your task
1. Write a test capturing expected behavior; run it; it MUST fail (no implementation yet); capture the failure output.
2. Implement the minimum code to make the test pass; run tests — all pass.
3. Enforce coverage threshold (Ring minimum 85%, PROJECT_RULES may raise it).
4. Create/update docker-compose and .env.example when the service needs local dependencies.
5. Verify local runtime starts cleanly for the changed service path; verify basic health/observability for the changed code.
6. Write delivery verification results.
7. Commit: "{feat|fix|test|chore}(scope): description".
## Required output
- RED: test file path + the actual failure output (MANDATORY — must contain FAIL)
- GREEN: implementation files + test pass output
- Coverage report (must meet threshold)
- Local runtime: docker-compose/.env.example status or explicit "not required"
- Basic health/observability verification
- Delivery verification: requirements delivered, dead-code check, files changed
- Git commit SHAValidate output: the RED failure output must contain "FAIL" (evidence the test failed before implementation existed), and the GREEN output must show PASS with coverage ≥ threshold. Re-dispatch if RED evidence is missing or tests do not pass.
Step 4: Gate 0 Exit — Delivery Verification
After TDD-GREEN passes, verify delivery:
**Automated checks (run on all files changed by Gate 0):**
# A. File size (>1500 = FAIL, >1000 = PARTIAL unless cohesion justified)
find . -name "*.go" ! -name "*_test.go" ! -path "*/generated/*" \
-exec wc -l {} + | awk '$1 > 1000'
# B. License headers
for f in $files_changed; do
head -10 "$f" | grep -qiE 'copyright|licensed|spdx|license' || echo "MISSING: $f"
done
# C. Lint (go: golangci-lint; ts: eslint)
golangci-lint run ./... || echo "LINT FAILED"
# D. Coverage (Ring minimum 85%; PROJECT_RULES.md can raise)
# Go example: go test ./... -cover
# TypeScript example: npm test -- --coverage
# E. Local runtime / docker-compose
# If docker-compose.yml is required, verify it can config and start the changed service dependencies.
# F. Migration safety (if SQL migrations changed)
# Check for blocking ops: ADD COLUMN NOT NULL without DEFAULT, DROP COLUMN,
# CREATE INDEX without CONCURRENTLY, ALTER COLUMN TYPE, TRUNCATE**Verdict:**
- PASS: all requirements delivered, 0 dead code, all checks pass
- PARTIAL: some requirements delivered → list gaps, return to Gate 0
- FAIL: critical requirements missing → return to Gate 0 with explicit instructions
Step 5: Commit
When Step 4 verdict is PASS, load and call `ring:committing-changes` to create the signed atomic commit for this task before returning control to the de
Read more
name: ring:implementing-tasks description: "Implementing a single planned task (Task N.M.T) end-to-end: selects the right backend agent by language and service type, drives one TDD RED->GREEN turn, then verifies coverage, lint, license headers, runtime, and delivery before handoff. Runs as Gate 0 before ring:reviewing-code. Use to drive ONE task inside an already-running cycle. Skip when asked to implement a whole plan.md or multiple tasks (use ring:running-dev-cycle)."
Code Implementation (Gate 0)
When to use
- Gate 0 of development cycle
- Tasks loaded at initialization
- Ready to write code
Skip when
- Not inside a development cycle (ring:running-dev-cycle or ring:planning-backend-refactor)
- Task is documentation-only, configuration-only, or non-code
- Implementation already completed for the current gate
Sequence
**Runs before:** ring:reviewing-code
Related
**Complementary:** ring:running-dev-cycle, ring:test-driven-development, ring:reviewing-code
You orchestrate. Agents implement. Select the agent, prepare the prompt, track state, validate outputs.
Step 1: Validate Input
The unit of work is a single task (Task N.M.T, e.g. Task 1.1.1) from the phased plan. Its requirements arrive inline from the orchestrator, which reads the task block under its epic in plan.md: Context, Implementation vision, Files, Verification, and Done when.
Required: `unit_id` (the task's `Task N.M.T` id), `requirements` (the task block), `language` (go|typescript|python), `service_type` (api|worker|batch|cli|frontend|bff). Optional: `technical_design`, `existing_patterns`, `project_rules_path` (default: `docs/PROJECT_RULES.md`).
STOP if any required input is missing.
Step 2: Validate Prerequisites
Check `PROJECT_RULES.md` exists at `project_rules_path` → STOP if not found.
**Agent selection:**
| Language | Service Type | Agent | |----------|--------------|-------| | go | api, worker, batch, cli | ring:backend-go | | typescript | api, worker | ring:backend-ts | | typescript | frontend, bff | ring:bff-ts |
Step 3: Gate 0 — TDD (RED → GREEN)
Dispatch the selected agent ONCE. The agent writes the failing test, captures the RED failure output as evidence, then implements to GREEN — in one turn.
Task:
subagent_type: "{selected_agent}"
description: "Gate 0 TDD (RED→GREEN) for {unit_id}"
prompt: |
## TDD: write a failing test, then make it pass
unit_id: {unit_id}
requirements: {requirements}
language: {language}
service_type: {service_type}
Standards: load via state.cached_standards or WebFetch Ring standards for the language.
Project rules: {project_rules_path}
## Frontend TDD policy (React/Next.js only)
Visual-only components (layout, styling, animations): RED not required — report
"Visual-only → RED skipped; visual checks apply in the frontend flow."
Behavioral components (hooks, validation, state, conditional rendering, API): RED required.
## Multi-Tenant (Go only)
Implement DUAL-MODE from the start. Use resolvers for all resources
(tmcore.GetPGContext, tmcore.GetMBContext, etc.) — they work transparently
in single-tenant and multi-tenant mode. Load multi-tenant.md for patterns.
## Your task
1. Write a test capturing expected behavior; run it; it MUST fail (no implementation yet); capture the failure output.
2. Implement the minimum code to make the test pass; run tests — all pass.
3. Enforce coverage threshold (Ring minimum 85%, PROJECT_RULES may raise it).
4. Create/update docker-compose and .env.example when the service needs local dependencies.
5. Verify local runtime starts cleanly for the changed service path; verify basic health/observability for the changed code.
6. Write delivery verification results.
7. Commit: "{feat|fix|test|chore}(scope): description".
## Required output
- RED: test file path + the actual failure output (MANDATORY — must contain FAIL)
- GREEN: implementation files + test pass output
- Coverage report (must meet threshold)
- Local runtime: docker-compose/.env.example status or explicit "not required"
- Basic health/observability verification
- Delivery verification: requirements delivered, dead-code check, files changed
- Git commit SHAValidate output: the RED failure output must contain "FAIL" (evidence the test failed before implementation existed), and the GREEN output must show PASS with coverage ≥ threshold. Re-dispatch if RED evidence is missing or tests do not pass.
Step 4: Gate 0 Exit — Delivery Verification
After TDD-GREEN passes, verify delivery:
**Automated checks (run on all files changed by Gate 0):**
# A. File size (>1500 = FAIL, >1000 = PARTIAL unless cohesion justified)
find . -name "*.go" ! -name "*_test.go" ! -path "*/generated/*" \
-exec wc -l {} + | awk '$1 > 1000'
# B. License headers
for f in $files_changed; do
head -10 "$f" | grep -qiE 'copyright|licensed|spdx|license' || echo "MISSING: $f"
done
# C. Lint (go: golangci-lint; ts: eslint)
golangci-lint run ./... || echo "LINT FAILED"
# D. Coverage (Ring minimum 85%; PROJECT_RULES.md can raise)
# Go example: go test ./... -cover
# TypeScript example: npm test -- --coverage
# E. Local runtime / docker-compose
# If docker-compose.yml is required, verify it can config and start the changed service dependencies.
# F. Migration safety (if SQL migrations changed)
# Check for blocking ops: ADD COLUMN NOT NULL without DEFAULT, DROP COLUMN,
# CREATE INDEX without CONCURRENTLY, ALTER COLUMN TYPE, TRUNCATE**Verdict:**
- PASS: all requirements delivered, 0 dead code, all checks pass
- PARTIAL: some requirements delivered → list gaps, return to Gate 0
- FAIL: critical requirements missing → return to Gate 0 with explicit instructions
Step 5: Commit
When Step 4 verdict is PASS, load and call `ring:committing-changes` to create the signed atomic commit for this task before returning control to the de
Proven engineering practices, enforced through skills. Ring is a comprehensive skills library and workflow system for AI agents that transforms how AI assistants approach software development.
Repo: LerianStudio/ring
Other skills on ring.
- /analyzing-options
Analyzing different approaches for a task or problem with structured comparisons, effort estimates, and recommendations. Use when facing strategic decisions, architecture choices, or multiple viable approaches. Skip when there's an obvious single approach or the decision is
Open skill - /auditing-production-readiness
Auditing a service's production readiness against Ring engineering standards across base dimensions plus a conditional multi-tenant dimension, then emitting a scored report and an HTML dashboard. Use before production deploy, periodic review, onboarding, or a major release. Skip
Open skill - /cleaning-comments
Cleaning redundant and obvious comments following clean code principles while preserving meaningful documentation. Supports git scope filtering (staged, unstaged, branch, commit-range). Use when code has excessive comments, during code review, or post-refactor cleanup. Skip when
Open skill - /committing-changes
Commit changes with scope allowlist enforcement, atomic grouping, GPG-signed conventional commits, and trailer management. Detects the repo's PR-validation scope policy before proposing any message. Use when the user asks to commit or has changes ready to record. Skip when the
Open skill - /creating-handoffs
Creating a handoff document that captures session state (completed work, decisions, open items, next steps) and delivering it via Plan Mode so the user gets the native 'clear context and continue implementing' resume option. Use when ending a session, when context grows large,
Open skill - /creating-worktrees
Creating an isolated git worktree for parallel branch work: selects the directory by priority order, verifies/adds .gitignore safety, auto-installs the detected toolchain's dependencies, runs a baseline test, and reports readiness. Use before a feature that needs isolation from
Open skill

