Task Continue Skill

Resume spec implementation from the last checkpoint. Picks up where task-dispatch left off.

When to Use

After hitting context limit during /implement
Starting a new session to continue a spec
Recovering from interruption mid-implementation

Workflow

Step 1: Find Checkpoint

Parse spec name from argument (e.g., /continue auth-system)
If no argument: find most recent checkpoint in ./specs/active/*/checkpoint.yaml
If no checkpoint found: suggest /implement instead

bash

# Find most recent checkpoint
ls -t ./specs/active/*/checkpoint.yaml | head -1

Step 2: Load Context

Read these files (in parallel):

./specs/active/<spec>/checkpoint.yaml  # Session state
./specs/active/<spec>/spec.md          # Requirements
./specs/active/<spec>/tasks.yaml       # Task definitions
./specs/active/<spec>/dependencies.yaml # Batch structure
./specs/active/<spec>/validation.yaml  # Review config

Step 3: Verify Branch State

bash

# Checkout spec branch
git checkout <checkpoint.branch>

# Verify at expected commit
git log -1 --format="%H" | head -c 8
# Should match checkpoint.last_commit

# If mismatch, warn user and ask to proceed or abort

Step 4: Report Progress

Present concise status:

## Resuming: <spec_name>

**Progress:** Batch <last_batch>/<total_batches> complete
**Completed:** <N> tasks
**Remaining:** <M> tasks

**Next batch:** #<next_batch.number>
- <task_id>: <task_name>
- <task_id>: <task_name>
[parallel: yes/no]

**Deferred issues:** <count>
[list if any]

Continuing with three-phase pipeline...

Step 5: Resume Three-Phase Pipeline

Execute the next batch using the same pipeline as task-dispatch:

Phase A: TESTERS
├── Dispatch task-tester(s) for next_batch.tasks
└── Wait for completion

Phase B: IMPLEMENTERS
├── Dispatch task-implementer(s) with tester reports
└── Wait for completion

Phase C: REVIEWERS
├── Dispatch 1 Claude + N OpenCode reviewers (from review_config)
└── Wait for completion + synthesize

CRITICAL: Follow all task-dispatch rules:

Always use model: opus for subagents
Dispatch parallel subagents in single message
Never skip reviewer phase
Fix Critical/High issues before proceeding

Step 6: Update Checkpoint

After batch completes:

Update tasks.yaml statuses
Write new checkpoint.yaml
Commit with batch info
Continue to next batch or report completion

Quick Resume Template

When resuming, use this condensed context for subagents:

For Tester:

Task: <task_id> - <task_name>
From: <spec_name> (batch <N>)
Requirements: [from tasks.yaml]
Test hints: [from tasks.yaml]

Invoke `code-test` skill. Write failing tests (RED).
Report tester_report YAML.

For Implementer:

Task: <task_id> - <task_name>
From: <spec_name> (batch <N>)
Tester report: [paste tester_report]

Invoke `code-implement` skill. Make tests pass (GREEN).
Report implementer_report YAML.

For Reviewer:

Batch <N> review for <spec_name>
Tasks: <task_ids>
Implementer reports: [paste all]
Spec requirements: [from tasks.yaml]

Invoke `code-review` skill. Evaluate gates.
Report reviewer_report YAML.

Handling Edge Cases

Checkpoint not found:

No checkpoint found for <spec>.
Run /implement <spec> to start fresh.

Branch mismatch:

Warning: Current branch differs from checkpoint.
Expected: feat/<spec> at <sha>
Actual: <current_branch> at <current_sha>

Options:
1. Checkout checkpoint branch and continue
2. Abort and investigate

Checkpoint stale (tasks.yaml modified):

Warning: tasks.yaml modified since checkpoint.
Checkpoint: <timestamp>
tasks.yaml: <modified_time>

Regenerating next batch from current state...

All tasks complete:

All tasks complete for <spec>.
Run final review? [Y/n]

Integration

Command: /continue [spec-name]

Related skills:

task-dispatch - Initial execution (writes checkpoints)
spec-update - Sync task status from git history
spec-archive - Archive completed spec

Example Session

User: /continue

Claude: Found checkpoint for auth-system

## Resuming: auth-system

**Progress:** Batch 2/4 complete
**Completed:** 4 tasks (T001-T004)
**Remaining:** 3 tasks

**Next batch:** #3
- T005: Add session management
- T006: Add token refresh
[parallel: yes]

**Deferred issues:** 1
- [M1] Variable naming in auth.py:45 (batch 2)

Continuing with three-phase pipeline...

[Dispatches 2 testers in parallel]
[Dispatches 2 implementers in parallel]
[Dispatches 1 Claude + 2 OpenCode reviewers in parallel]
[Synthesizes review, no critical issues]
[Writes checkpoint, commits batch 3]

Batch 3 complete. 1 batch remaining.
Continue in this session or /continue later.

Search AI Tools

task-continue

Install this agent skill to your Project

SKILL.md