full-output-enforcement

Enforces complete, unabridged output from coding agents — bans placeholder patterns like "// rest of code" or "TODO," and handles token-limit cutoffs with a clean resume mechanism instead of silent truncation.

LeonxlnxTop 10 reviews

Skill description

Output only the complete code

Forces coding agents to deliver complete, unabridged output. This skill bans placeholder patterns like "// rest of code" or "TODO," and hedging phrases like "I can provide more details if needed." When a task is too long for one response, the agent stops at a clean breakpoint and lets the user say "continue" to resume exactly where it left off, instead of silently truncating.

insights

Bans Placeholder Patterns

Hard-blocks lazy code comments like "// ...", "// implement here", "// TODO," and prose hedges like "the rest follows the same pattern." The agent must write real, runnable code — not a description of what the code would do.

psychology

Clean Token-Limit Handling

When a response hits the token limit, the agent stops at a clean breakpoint (end of a function, file, or section) and prints "[PAUSED — X of Y complete]," letting the user reply "continue" to resume exactly where it stopped — no recap, no repetition.

Installation guide

Set up MCP integration

Connect the skill to your agent environment with the steps below.

1
Go to Dashboard -> MCP Key, name your key, create it, and copy it immediately. The key is shown only once.
Go to Dashboard -> MCP Key, name your key, create it, and copy it immediately. The key is shown only once.
2
Add the MCP configuration below to Claude Code, Cursor, Antigravity, or your IDE. Replace YOUR_MCP_KEY with the key from step 1.
Add the MCP configuration below to Claude Code, Cursor, Antigravity, or your IDE. Replace YOUR_MCP_KEY with the key from step 1.
3
Restart the IDE or reload MCP servers. Then ask your agent to use the skill with the prompt below.
Restart the IDE or reload MCP servers. Then ask your agent to use the skill with the prompt below.

mcp.json

{
  "mcpServers": {
    "nskill": {
      "serverUrl": "https://[your-project].supabase.co/functions/v1/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_MCP_KEY"
      }
    }
  }
}

Use skill [tên-skill] from nskill to do this: [describe your request]

Frequently asked questions

Does this skill slow down response speed?expand_more

It can, since the agent must generate everything in full instead of shortening it. In exchange, you get output that's usable right away instead of something you have to finish yourself.

What happens if a task is genuinely too long for one response?expand_more

The agent stops at a clean breakpoint (end of a function, file, or section) and prints "[PAUSED — X of Y complete]." Just reply "continue" and it resumes exactly where it left off — no recap, no repetition.

Does this skill apply to non-code tasks too?expand_more

Yes. The ban on hedging phrases like "the rest follows the same pattern" or "I can add more if needed" applies to any text-based output, not just code.

Reviews (0)

0.0

starstarstarstarstar

Based on 0 reviews

5star0

4star0

3star0

2star0

1star0

Write a review

Share your experience with this skill to help other users.

No reviews yet. Be the first to share your experience.