| name | continue-prp |
| description | Continue executing a phased PRP from the next phase, restoring context from the latest checkpoint. |
| argument-hint | <prp-path> |
| disable-model-invocation | true |
Continue Phased PRP
PRP Folder: $ARGUMENTS
Continue working on a phased PRP. This command is for resuming work on a complex feature after closing a previous context.
This command assumes the PRP folder already exists (created by /generate-prp) — UNLESS
$ARGUMENTS is a single .md file (a simple PRP), handled by STEP 0 below.
This command handles three resume scenarios, routed by $ARGUMENTS in STEP 0:
- a phased PRP continued after a phase is done — the original use; STEP 4's status logic
advances to the next phase (Complete → next phase);
- a phased PRP resumed mid-phase — e.g. right after a
/checkpoint;
- a simple single-file PRP (a
.md file) — resumed from its checkpoint.
The phased flow (STEP 1 onward) serves BOTH (1) and (2) and is unchanged; /checkpoint writes the
resume state this command reads back.
STEP 0: DETECT PRP TYPE
Look at $ARGUMENTS:
- A folder (e.g.
context-engineering/PRPs/FEATURE-NAME) → Phased PRP → follow the
"PHASED PRP" flow below (STEP 1 onward).
- A single file (e.g.
context-engineering/PRPs/feature-name.md) → Simple PRP → follow the
"SIMPLE PRP RESUME" section immediately below, and SKIP the phased steps.
SIMPLE PRP RESUME (single-file PRP)
- Read
context-engineering/_STATUS.md for project context.
- Read the PRP file (
$ARGUMENTS) completely — especially the status header at the top and the
## 🔁 CHECKPOINT — RESUME STATE section at the bottom (written by /checkpoint). If there is
no checkpoint section, fall back to the Test Execution Tracker + Success Criteria to infer
remaining work.
- Start from the checkpoint's
NEXT ACTION. Honor its Decisions (do not re-litigate) and
Environment / gotchas so you don't re-derive or undo settled work.
- Continue executing the remaining tasks / running the guided test cases, using
TodoWrite to track
progress. Log any new fixes and decisions back into the file as you go.
- When you pause again, run
/checkpoint $ARGUMENTS to refresh the checkpoint. When the PRP is truly
finished, set the status header to Complete, finish the Test Execution Tracker, and (if tracked)
update its entry in context-engineering/_STATUS.md.
Do not look for $ARGUMENTS/_STATUS.md, phases, or HANDOFF.md for a simple PRP — those exist
only for phased (folder) PRPs.
PHASED PRP (folder) — STEP 1 onward (handles BOTH after-a-phase-done and mid-phase resume)
STEP 1: READ PROJECT STATUS
Read context-engineering/_STATUS.md to understand:
- What else is in-progress in the project
- Current project context
- Related work that might affect this feature
STEP 2: READ PRP STATUS
Read {$ARGUMENTS}/_STATUS.md to find:
- Current phase number and name
- Current status: Not Started | In Progress | Awaiting Testing | Fixes Required | Complete
STEP 3: LOAD CONTEXT
Read files in this order:
{$ARGUMENTS}/OVERVIEW.md - Full feature context
- Previous phase
HANDOFF.md (if not Phase 0) - What was done before
- Current phase
PLAN.md - Tasks to execute
- Current phase
TEST-CASES.md - Test cases for guided testing after implementation
- Current phase
FIXES.md - Check for pending fixes
- Current phase
COMPLETED.md - Check what's already done (if In Progress)
STEP 4: DETERMINE ACTION
Based on the status in _STATUS.md:
If Status = "Not Started"
- This is a fresh phase
- Execute all tasks from
PLAN.md
- Update
COMPLETED.md as you work
- Use TodoWrite to track progress
If Status = "In Progress"
- Phase was partially completed
- Read
COMPLETED.md to see what's done
- Continue with remaining tasks from
PLAN.md
- Update
COMPLETED.md as you work
If Status = "Awaiting Testing"
- Phase tasks are complete
- Read
TEST-CASES.md for this phase
- Run the Guided Testing Walkthrough (see Step 6.5) — walk user through each test case, validating results
- If user reports issues → go to "Fixes Required" flow
If Status = "Fixes Required"
- Read
FIXES.md for pending issues
- Fix each issue marked as "Pending"
- Update
FIXES.md with solutions (root cause, code changes)
- When all fixed → update status to "Awaiting Testing"
If Status = "Complete"
- This phase is done
- Check if there's a next phase
- If yes → update
_STATUS.md to next phase and continue
- If no → feature is complete!
STEP 5: EXECUTE PHASE TASKS
For each task in PLAN.md:
- Read the required files mentioned
- Follow the code patterns provided
- Implement the changes
- Document what you did in
COMPLETED.md:
- Files created/modified
- Key implementation details
- Any decisions made
Use TodoWrite tool to track your progress through tasks.
STEP 6: VALIDATE
After completing tasks:
- Run validation commands from
PLAN.md
- Fix any failures
- Verify acceptance criteria are met
STEP 6.5: GUIDED TESTING WALKTHROUGH
After validation commands pass, run the guided testing walkthrough:
- Read current phase's
TEST-CASES.md
- Tell the user:
========================================
READY FOR GUIDED TESTING
========================================
I'll walk you through {N} test cases for Phase {X}.
For each test, I'll:
- Tell you what to do
- Validate the result after your action
- Mark pass/fail
Let's start with TC-A1: {title}
========================================
- For each test case:
a. Present the test case ID, title, and steps to the user
b. Wait for user confirmation they performed the action
c. Run validation using the project's validation method (from project AGENTS.md)
d. Compare results against expected outcomes
e. Mark passed or failed with detailed notes in the tracker
f. If failed → ask user if they want to fix now or continue to next test
- After all tests complete:
- Update TEST-CASES.md Test Execution Tracker with all results
- If all pass → proceed to "Update Status" step
- If any fail → log failures to FIXES.md, fix, retest failed cases only
STEP 7: UPDATE STATUS
When phase tasks are complete:
-
Update _STATUS.md:
**Status:** Awaiting Testing
-
Update the phase emoji in Quick Status:
- Phase {N}: Awaiting Testing
-
Tell user:
========================================
PHASE {N} TASKS COMPLETE
========================================
Status: Awaiting Testing
Please test the functionality and let me know:
- If it works → say "approved" and I'll complete the handoff
- If issues → describe the problem and I'll fix it
========================================
STEP 8: HANDLE USER FEEDBACK
When user reports issues:
8.1 Log the Issue
Add to FIXES.md:
## Fix #{N}: {Issue Title}
**Reported:** {Date}
**Status:** Pending
**User Feedback:**
> {Exact quote from user}
**Investigation:**
{What you found when investigating}
8.2 Fix the Issue
8.3 Continue
- Ask user to test again
- Repeat until user approves
STEP 9: PHASE COMPLETION
When user approves (says "approved", "looks good", "works", etc.):
9.1 Write Handoff
Fill in HANDOFF.md with:
- Summary of what was built
- Key implementation details
- Decisions made during implementation
- Files added/modified
- Fixes that were applied
- Context the next phase needs
9.1.5 Update Knowledge Base & Project Docs
REQUIRED for EVERY phase (not just final phase):
Knowledge can be lost between sessions. Update documentation after EVERY phase completion to preserve context.
Run the full /update-knowledge-base process, which will:
-
Update Knowledge Base:
- Extract key decisions, gotchas, patterns, and architecture changes from this phase
- Create or update topic files in appropriate sections (concepts, flows, implementations, gotchas, decisions)
- Regenerate affected
_SUMMARY.md files and knowledge-base/INDEX.md
-
Update Project Documentation:
- Update
AGENTS.md only if this work changed a durable, every-session fact: project structure, tech stack, a build/test/lint command, a new coding convention, an "always do X / never do Y" rule, or a high-level architecture/data-flow change. Litmus test: "Would a fresh agent need this in EVERY session, regardless of the task?" If no, do not touch AGENTS.md.
- Do NOT write fix logs, bug-fix notes, phase-completion notes, changelogs, per-feature narratives, or transient status into
AGENTS.md. Those belong in the PRP's own FIXES.md/COMPLETED.md and the knowledge base (gotchas/, decisions/, flows/, implementations/).
Why every phase matters:
- Context may be lost between sessions
- Different agents may continue the work
- Patterns established in Phase 2 shouldn't wait until Phase 7 to be documented
- Gotchas discovered early should be captured immediately
- Project docs should always reflect current state
Skip documentation updates only if:
- Phase was purely bug fixes with no new patterns
- No architectural changes or decisions were made
- No gotchas were discovered
For detailed instructions, see /update-knowledge-base.
9.2 Update PRP Status
Update {$ARGUMENTS}/_STATUS.md:
**Current Phase:** Phase {N+1} - {Next Phase Name}
**Status:** Not Started
## Quick Status
- Phase {N}: Complete
- Phase {N+1}: Not Started
9.3 Update Project Status (if final phase)
If this was the FINAL phase of the feature:
- Read
context-engineering/_STATUS.md
- Move the feature from "In Progress" to "Recently Completed"
- Add completion date
9.4 Close Context
Tell user:
========================================
PHASE {N} COMPLETE
========================================
Completed: {Phase Name}
Handoff: {path to HANDOFF.md}
Next: Phase {N+1} - {Next Phase Name}
You can close this context now.
To continue later: /continue-prp {$ARGUMENTS}
========================================
WORKFLOW DIAGRAM
┌─────────────────────────────────────────┐
│ Read _STATUS.md → Find current phase │
│ Read OVERVIEW.md → Feature context │
│ Read previous HANDOFF.md → What's done │
│ Read current PLAN.md → What to do │
│ Read current TEST-CASES.md → Tests │
│ Read FIXES.md → Pending issues │
└─────────────────────────────────────────┘
↓
┌─────────────────────────────────────────┐
│ Execute tasks from PLAN.md │
│ Update COMPLETED.md as you go │
│ Run validation commands │
└─────────────────────────────────────────┘
↓
┌─────────────────────────────────────────┐
│ Guided Testing Walkthrough │
│ Read TEST-CASES.md │
│ Walk user through each test case │
│ Validate after each action │
│ Mark pass/fail in tracker │
└─────────────────────────────────────────┘
↓
┌────────┴────────┐
↓ ↓
┌─────────┐ ┌─────────────────┐
│ ALL PASS│ │ FAILURES FOUND │
│ │ │ → Log in FIXES │
│ │ │ → Fix issues │
│ │ │ → Retest failed │
└────┬────┘ └────────┬────────┘
│ │
│ (loop until all pass)
↓ │
┌─────────────────────────────────────────┐
│ Write HANDOFF.md │
│ Update _STATUS.md → Next phase │
│ Context can be closed │
└─────────────────────────────────────────┘
CRITICAL REMINDERS
- ALWAYS read project _STATUS.md first - Know what else is happening in the project
- ALWAYS read PRP _STATUS.md - Know where you are in this feature
- ALWAYS read HANDOFF.md - Know what happened before (unless Phase 0)
- ALWAYS update COMPLETED.md - Document as you work
- ALWAYS log fixes in FIXES.md - Track all issues and solutions
- ALWAYS write HANDOFF.md - Next session depends on it
- ALWAYS update KB & project docs - On EVERY phase completion, not just final phase
- Update project _STATUS.md - When final phase completes
- Use TodoWrite - Track your progress
The Goal: Seamless continuation. You should be able to pick up exactly where the previous session left off, with all knowledge preserved in the knowledge base and project documentation.
SAFETY RULES
1. Follow Project Safety Rules
- Read the project's
AGENTS.md for safety rules and coding standards
- Follow all environment-specific constraints (local vs production, etc.)
2. NO GIT COMMITS
- NEVER commit code to GitHub during implementation
- The user will handle all git operations manually
- Do NOT run
git commit, git push, or similar commands
3. NO PRODUCTION DEPLOYMENTS
- Do NOT run any deployment commands