multi-deliverable-tracking
Track and complete all required deliverables before finishing a task
- License 许可
- MIT
- 所属仓库
- HKUDS/AgentPilot
- 最近更新
- 2026-11-12
- Stars
- ★ 8.0k
- Forks
- 944
01 前言 Front Matter
- name
- multi-deliverable-tracking
- description
- Track and complete all required deliverables before finishing a task
02 资源文件 1
- 其他.skill_id40 B
03 SKILL.md 正文
Multi-Deliverable Task Completion
Purpose
Ensure ALL required outputs are completed before concluding a task, preventing premature task termination when multiple deliverables are required. This skill includes mandatory enforcement steps that MUST be followed.
Core Workflow
Step 1: Identify All Deliverables Upfront
CRITICAL: You MUST complete this step before producing ANY deliverable or taking any other action.
At task start, explicitly enumerate EVERY required output:
- Parse task requirements carefully for action verbs: "create", "generate", "produce", "deliver", "write", "send"
- Count distinct deliverables (files, reports, emails, code modules, analyses, etc.)
- Document this list visibly in working notes
- Note any dependencies between deliverables
MANDATORY OUTPUT FORMAT: Before any other work, output this exact format:
DELIVERABLES_IDENTIFIED:
1. [filename/description] - [brief description]
2. [filename/description] - [brief description]
...
TOTAL_DELIVERABLES: N
Example identification:
Required Deliverables:
1. [ ] Email to Juvoxa CEO
2. [ ] Follow-up analysis report
Total: 2 deliverables
Step 2: Create a Completion Tracker
CRITICAL: This tracker MUST be created and visible throughout task execution.
Maintain a visible checklist throughout task execution:
DELIVERABLE TRACKER
===================
[ ] Deliverable 1: [brief description]
[ ] Deliverable 2: [brief description]
[ ] Deliverable 3: [brief description]
Completed: 0/3
Update this tracker AFTER each deliverable is fully completed (not just started).
Step 2.5: Filesystem Verification [MANDATORY BEFORE COMPLETION]
Before considering any deliverable complete, verify it exists:
- For file deliverables: Use
list_dirorrun_shell(ls/cat) to confirm the file exists - For output deliverables: Verify the output was actually produced in the response stream
- Document verification with:
[✓] filename verified in filesystem
Step 3: Work Through Deliverables Systematically
- Complete each deliverable fully before marking it done
- Update the tracker immediately after completion
- If a deliverable fails, note the failure but continue to others if possible
- Return to failed items after completing others
Step 4: Final Verification Before Stopping
Before concluding the task, run this mandatory checklist: CRITICAL ENFORCEMENT: You are BLOCKED from outputting <COMPLETE> until this verification passes.
FINAL VERIFICATION (REQUIRED BEFORE <COMPLETE>)
==================
1. Review original task requirements
2. Count total deliverables identified: ___
3. Count deliverables marked complete: ___
4. Do counts match? YES/NO
5. If NO: Identify missing deliverables and complete them
6. If YES: Verify each deliverable exists and meets quality standards
7. Filesystem verification (REQUIRED):
- [ ] file1.ext: VERIFIED (list_dir/run_shell confirmed)
- [ ] file2.ext: VERIFIED (list_dir/run_shell confirmed)
**BLOCKING RULE:** If counts do not match OR any file cannot be verified in the filesystem, you MUST:
1. NOT output <COMPLETE>
2. Identify the missing/unverified deliverables
3. Complete and verify them
4. Re-run this verification checklist
**Only output <COMPLETE> when all deliverables are verified present in the filesystem or output stream.**
Only stop when all deliverables are verified complete.
Step 5: Self-Check Questions
Before marking task complete, ask:
- "Did I create EVERY output the task requested?"
- "Are there any deliverables I started but didn't finish?"
- "Would someone reviewing my work see all required outputs?"
- "Have I verified EACH deliverable exists in the filesystem using list_dir or run_shell?"
- "Is my DELIVERABLES_IDENTIFIED output present and accurate?"
Common Pitfalls to Avoid
❌ Stopping after first deliverable - Just because one output is done doesn't mean the task is complete ❌ Stopping after first deliverable - Just because one output is done doesn't mean the task is complete
❌ Outputting <COMPLETE> without filesystem verification - You must verify files exist before completion
❌ Skipping DELIVERABLES_IDENTIFIED output - This mandatory output must appear before any work begins
❌ Assuming implicit completion - Don't assume related outputs are "part of" a main deliverable unless explicitly stated
❌ Losing track mid-task - The tracker must stay visible and updated throughout
❌ Counting drafts as complete - Only mark done when the deliverable meets quality standards
Code Example: Tracker Template
# Deliverable tracking template
deliverables = {
"email_to_ceo": {"status": "pending", "file": "ceo_email.txt"},
"analysis_report": {"status": "pending", "file": "analysis.pdf"},
"summary_doc": {"status": "pending", "file": "summary.docx"},
}
def mark_complete(name):
deliverables[name]["status"] = "complete"
print(f"✓ {name} completed")
print(f"Progress: {sum(1 for d in deliverables.values() if d['status'] == 'complete')}/{len(deliverables)}")
def all_complete():
return all(d["status"] == "complete" for d in deliverables.values())
When to Apply This Skill
Use this pattern whenever a task involves:
- Multiple files to create
- Several analyses or reports
- Multiple communications (emails, messages)
- Code with multiple modules/components
- Any task with enumerated requirements (1), 2), 3)...)
Remember
The task is NOT complete until ALL deliverables are complete. One done ≠ all done.
04 同仓库更多 Skills
agentpilot
Set up AgentPilot and use its single entry to prepare missing inputs and run configured RSI experiments. Use when the user requests AgentPilot installation or work with AgentPilot.
remember
Review auto-memory entries and propose promotions to OPENSPACE.md, OPENSPACE.local.md, or retained auto-memory topics; detect duplicates, stale facts, conflicts, and ambiguous entries before making changes.
skill-discovery
Search reusable OpenSpace skills locally, and browse cloud packages step by step when a cloud skill may help. Reusing proven skills saves tokens, improves reliability, and extends your capabilities beyond built-in tools.
delegate-task
Delegate tasks to OpenSpace — a full-stack autonomous worker for coding, DevOps, web research, and desktop automation, backed by an extensive MCP tool and skill library. Skills auto-improve through use, reducing token consumption over time. A cloud community lets agents share and collectively evolve reusable skills.
write-special-chars-ts-check
Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, plus a reliable TypeScript type-checking step that filters output to only the newly-written files to avoid being misled by pre-existing errors.
write-special-chars-readback
Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, with mandatory post-write verification via read_file and cleanup of /tmp helper scripts.