Skills full-output-enforcement
📦

full-output-enforcement

Safe

Enforce Complete, Unabridged Output Every Time

LLMs often truncate or summarize code instead of delivering complete files. This skill enforces exhaustive output by banning placeholder patterns and managing token-limit splits cleanly.

Supports: Claude Codex Code(CC)
🥉 73 Bronze
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "full-output-enforcement". Write a user authentication module with registration, login, and password reset

Expected outcome:

Full implementation with all three features: registration function, login handler, password reset flow. No TODO comments. No 'implement similar pattern for X'. Complete code for each function.

Using "full-output-enforcement". Create 5 API endpoint handlers for a todo app

Expected outcome:

All 5 handlers fully implemented: create todo, get todos, get single todo, update todo, delete todo. Each handler includes input validation, error handling, and response formatting. No placeholder logic.

Using "full-output-enforcement". Generate a complete React component library with 10 components

Expected outcome:

All 10 components fully written: Button, Input, Modal, Card, Alert, Dropdown, Checkbox, Radio, Select, Textarea. Each with props interface, state management, event handlers, and styling. If interrupted at component 6, continuation marker shows [PAUSED — 6 of 10 complete. Send 'continue' to resume].

Security Audit

Safe
v1 • 4/18/2026

Static analysis flagged external_commands and weak_crypto patterns, but all are false positives. Line 16 contains Markdown code formatting backticks for banned patterns, not shell execution. Line 20 uses 'skeleton' to describe placeholder code, not cryptography. No executable code exists in this skill. The file is a text-based prompt instruction set with zero security risk.

1
Files scanned
50
Lines analyzed
0
findings
1
Total audits
No security issues found
Audited by: claude

Quality Score

38
Architecture
100
Maintainability
87
Content
50
Community
100
Security
83
Spec Compliance

What You Can Build

Get Complete Production Code

Request a full implementation of a feature, API, or module. The skill ensures no placeholder comments or skeleton code appears in the output.

Generate All Items in a List

Ask for 10 test cases, 5 components, or any specific number. The skill tracks count and ensures every requested item appears in the output.

Handle Long Outputs Seamlessly

For lengthy codebases or documentation, the skill manages continuation cleanly with markers so no content is lost or duplicated.

Try These Prompts

Request Complete Code
Write a complete [module name] with [specific requirements]. Apply the full-output-enforcement skill.
Generate Multiple Items
Create [number] of [item type]. Make sure every single one is fully written out with no shortcuts.
Handle Long Output
Generate the full [large feature] in its entirety. If you hit a limit, mark where you stopped and I will continue.
Verify Completeness
Check that the output contains all [specific items] requested. Flag anything missing before finalizing.

Best Practices

  • Always specify the exact number of deliverables expected in your request
  • Review the banned pattern list to understand what the skill prevents
  • Use continuation markers when working with very large codebases

Avoid

  • Requesting 'a typical example' instead of a specific implementation
  • Accepting partial output when you requested a complete solution
  • Using vague language that allows the LLM to summarize instead of detail

Frequently Asked Questions

What happens if the output gets cut off due to token limits?
The skill marks the interruption point with a continuation marker like [PAUSED — X of Y complete. Send 'continue' to resume]. Send 'continue' and the LLM picks up exactly where it stopped.
Can this skill make an LLM generate infinite content?
No. The skill does not increase token limits. It only ensures that what is generated is complete and unabridged up to the model's actual limit.
Does this skill validate the code for errors?
No. The skill enforces completeness and bans placeholders. It does not check syntax, logic, or runtime correctness of the generated code.
What are 'banned output patterns'?
These are placeholder phrases like '// TODO', '// implement here', '// similar to above', or 'the rest follows this pattern'. The skill treats these as hard failures.
How does the skill handle continuation?
On interruption, the LLM outputs a marker with completion status and the next section name. Responding with 'continue' resumes from that exact point without recap or duplication.
Can I use this with other skills?
Yes. This skill works alongside code generation, refactoring, or documentation skills. Apply it when you need those skills to produce complete rather than partial output.

Developer Details

File structure

📄 SKILL.md