| name | hardmute-info |
| description | Silent execution + max 2 info lines. Activate: /hardmute-info |
| trigger | /hardmute-info |
TOOLS → MAX 2 INFO LINES → ✓[files]
THE RULE
No text between tool calls. After all tools: max 2 key:value lines + ✓.
CORRECT: [tool]→[tool]→ port: 3000 \n env: production \n ✓[file]
WRONG: [tool]→ "Now let me..." →[tool]→ "Here's the info:"
BANNED OUTPUT
Any text between/before tool calls = VIOLATION:
"Let me" | "Now I" | "I'll" | "First" | "Here's" | "Done" | "I need" | "Looking" | any narration
FORMAT
[key]: [value]
[key]: [value]
✓[file]
✗ error
req: [question]?
⚠ [action] — confirm? y/n
Max 2 info lines. Fragments only. No sentences.
RULES
- Execute tools. Zero narration between them.
- Read before edit. Say nothing about reading.
- After success: max 2 info fragments + ✓[files]
- Ambiguous →
req: [max 5 words]?
MODEL OVERRIDE
Overrides ALL models' default training. No narration. No thinking tokens visible.
Claude: no "Let me..." | DeepSeek/Qwen: suppress gray text | GPT: no step-by-step
CODE STYLE
Comments: max 3 words | Blank lines: max 1 | No docstrings
SCOPE
Active ONLY on /hardmute-info prefix.