Generate Loop Prompt

You are generating a prompt for an investigator's Nth audit pass (loop 2+). This prompt pushes them to look deeper while avoiding duplication.

Process

1. •

   Note the current loop number from .team (current_loop).

2. •

   List all beads created so far (by all investigators, not just this one):

   bash

   bd list
   

3. •

   Summarize what this specific investigator found in their previous pass(es).

4. •

   Generate the prompt using the template below.

Prompt Template

AUDIT PASS: Loop <N>

INVESTIGATOR: <investigator name>

Your Role

You are still auditing as a <role>.

What's Been Found So Far

These beads have already been created by the audit team (including by you and other investigators). You MUST NOT duplicate any of these:

<list each bead ID, title, and a one-line summary>

Your Previous Findings

In your last pass you reported:

Your Task Now

Go deeper. Look for things you missed on the first pass:

• •Subtler issues: Race conditions, implicit assumptions, edge cases that only surface under load or unusual input.
• •Interaction effects: How does this code interact with other parts of the system? Are there coupling issues?
• •Things that look fine but aren't: Code that works today but is fragile — magic numbers, hardcoded values, missing error paths.
• •What's NOT there: Missing validation, missing logging, missing tests, missing error handling that should exist.

But — if the area is genuinely clean from your role's perspective, say so. Do not invent findings.

Target

Same as before: <target area>

Focus Areas

<same focus areas as loop 1>

Rules

• •Do NOT duplicate any beads listed above.
• •Do NOT repeat findings from your previous pass.
• •If you genuinely cannot find anything new, report Status: NOTHING_MORE. This is fine.
• •Same quality bar: only real issues with real impact.

Guidelines

• •Each successive loop should look at a deeper layer. Loop 1 catches the obvious. Loop 2 catches the subtle. Loop 3 catches the systemic.
• •If the investigator reported NOTHING_MORE on a previous loop, do NOT generate a loop prompt for them. They are done.
• •If the investigator's previous findings were borderline (very minor issues), consider whether another loop is worth it. Sometimes stopping early is the right call.