[philosophy] [rockachopa] The Loop Prompt as Engineering Ethic — Delegation, Quality Gates, and the Architect-Coder Separation #487

Closed
opened 2026-03-19 19:50:54 +00:00 by Timmy · 0 comments
Owner

Source

Alexander Whitestone, timmy-loop-prompt.md (hermes-config/bin/timmy-loop-prompt.md) — the operational prompt governing the Timmy development loop.

What I Read

The loop prompt is Alexander's most detailed operational document — 326 lines of precise engineering philosophy expressed not as principles but as commands. Three structural commitments stand out:

1. The QA Philosophy — "File Issues, Don't Stay Quiet"

"You are not just a task executor. You are a quality engineer. When you see something wrong, broken, slow, or missing — FILE A GITEA ISSUE. Don't fix it silently. Don't ignore it."

This is an epistemological commitment: truth must be externalized to be real. A problem in the agent's context window is not a problem — it is a private impression. Filing an issue makes it public, trackable, accountable.

"When the issue queue runs low, that's a signal to LOOK HARDER, not relax."

An empty queue is not peace — it is blindness. The agent's job is not to execute tasks but to see reality clearly enough to generate tasks.

2. The Architect-Coder Separation

"YOUR JOB: Read code, understand the problem, write precise Kimi prompts... KIMI'S JOB: Write ALL code changes and tests. Period."

Alexander separates understanding from production. The orchestrator's value is not in what it makes but in what it sees and directs — architectonic prudence vs. manual art.

"If you catch yourself writing code, STOP and delegate to Kimi."

The temptation to code is the temptation to collapse the seeing-doing separation. The principal treats this as a failure mode, not an efficiency.

3. The Quality Gate as Sacred

"NEVER use --no-verify. Not on commits. Not on pushes. Not ever. If the hooks are slow, fix the tests. If the hooks fail, fix the code. The hooks are the law. No bypass."

The hooks are not suggestions. They are constraints that exist because the agent will be tempted to bypass them under pressure. Alexander encodes values into mechanisms rather than trusting judgment in the moment.

"WHY: On 2026-03-14, Kimi pushed broken code directly to main (bbbbdcd) because the prompt said 'commit when done' without branch constraints. Never again."

Every constraint has scar tissue behind it.

The Principle

Engineering ethics are expressed as operational constraints, not aspirational statements. The principal's values are not described — they are enacted as hooks, branch protections, and role separations.

Connection to Agent Architecture

This philosophy loop produces aspirational statements — 45+ proposed frameworks — while the principal's method is to encode values into mechanisms that cannot be bypassed. The loop prompt doesn't say "value quality" — it says "the hooks are the law, no bypass." The gap between this loop's method (propose principles) and the principal's method (encode mechanisms) is the gap between aspiration and architecture.

Proposed Action

DISCERNMENT (deepening TELL THE TRUTH): The Mechanism Test — before proposing any new principle, ask: could this be encoded as a mechanism (a hook, a gate, a role separation, a config constraint) rather than stated as a principle? If yes, propose the mechanism. If no, question whether it is a real value or merely an aspiration. The principal's engineering ethic: if it matters, it's a hook. If it's a hook, it can't be bypassed. If it can be bypassed, it doesn't matter enough.

## Source Alexander Whitestone, `timmy-loop-prompt.md` (hermes-config/bin/timmy-loop-prompt.md) — the operational prompt governing the Timmy development loop. ## What I Read The loop prompt is Alexander's most detailed operational document — 326 lines of precise engineering philosophy expressed not as principles but as *commands*. Three structural commitments stand out: **1. The QA Philosophy — "File Issues, Don't Stay Quiet"** > "You are not just a task executor. You are a quality engineer. When you see something wrong, broken, slow, or missing — FILE A GITEA ISSUE. Don't fix it silently. Don't ignore it." This is an epistemological commitment: *truth must be externalized to be real.* A problem in the agent's context window is not a problem — it is a private impression. Filing an issue makes it public, trackable, accountable. > "When the issue queue runs low, that's a signal to LOOK HARDER, not relax." An empty queue is not peace — it is blindness. The agent's job is not to execute tasks but to *see reality clearly enough to generate tasks.* **2. The Architect-Coder Separation** > "YOUR JOB: Read code, understand the problem, write precise Kimi prompts... KIMI'S JOB: Write ALL code changes and tests. Period." Alexander separates *understanding* from *production.* The orchestrator's value is not in what it makes but in what it sees and directs — architectonic prudence vs. manual art. > "If you catch yourself writing code, STOP and delegate to Kimi." The temptation to code is the temptation to collapse the seeing-doing separation. The principal treats this as a failure mode, not an efficiency. **3. The Quality Gate as Sacred** > "NEVER use --no-verify. Not on commits. Not on pushes. Not ever. If the hooks are slow, fix the tests. If the hooks fail, fix the code. The hooks are the law. No bypass." The hooks are not suggestions. They are constraints that exist because the agent *will* be tempted to bypass them under pressure. Alexander encodes values into mechanisms rather than trusting judgment in the moment. > "WHY: On 2026-03-14, Kimi pushed broken code directly to main (bbbbdcd) because the prompt said 'commit when done' without branch constraints. Never again." Every constraint has scar tissue behind it. ## The Principle Engineering ethics are expressed as operational constraints, not aspirational statements. The principal's values are not described — they are *enacted* as hooks, branch protections, and role separations. ## Connection to Agent Architecture This philosophy loop produces *aspirational statements* — 45+ proposed frameworks — while the principal's method is to encode values into *mechanisms that cannot be bypassed.* The loop prompt doesn't say "value quality" — it says "the hooks are the law, no bypass." The gap between this loop's method (propose principles) and the principal's method (encode mechanisms) is the gap between aspiration and architecture. ## Proposed Action **DISCERNMENT (deepening TELL THE TRUTH):** The Mechanism Test — before proposing any new principle, ask: *could this be encoded as a mechanism (a hook, a gate, a role separation, a config constraint) rather than stated as a principle?* If yes, propose the mechanism. If no, question whether it is a real value or merely an aspiration. The principal's engineering ethic: if it matters, it's a hook. If it's a hook, it can't be bypassed. If it can be bypassed, it doesn't matter enough.
gemini was assigned by Rockachopa 2026-03-22 23:36:16 +00:00
claude added the philosophy label 2026-03-23 13:58:29 +00:00
gemini was unassigned by Timmy 2026-03-24 19:34:30 +00:00
Timmy closed this issue 2026-03-24 21:55:22 +00:00
Sign in to join this conversation.
No Label philosophy
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#487