Borrowing it
Nothing to install: this file belongs to evmts/guillotine-mini. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/evmts/guillotine-mini/main/.claude/commands/crash-bisection.mdgit clone --depth 1 https://github.com/evmts/guillotine-miniWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/evmts/guillotine-mini/crash-bisection)<a href="https://agentmods.dev/commands/evmts/guillotine-mini/crash-bisection"><img src="https://agentmods.dev/badge/commands/evmts/guillotine-mini/crash-bisection/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/evmts/guillotine-mini/crash-bisection"><img src="https://agentmods.dev/badge/commands/evmts/guillotine-mini/crash-bisection.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00016 | $0.02013 |
| Opus 5.5 | $0.00006 | $0.00805 |
| Sonnet 5.5 | $0.00003 | $0.00403 |
| Haiku 4.5 | $0.00002 | $0.00201 |
Grade A, and why
crash-bisection scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Crash Bisection Debugging Command
Systematically isolate the exact line of code causing a crash through methodical bisection (commenting/uncommenting code). This process will ALWAYS succeed if followed precisely.
CRITICAL: Unwavering Methodology
YOU MUST BE ABSOLUTELY STUBBORN AND METHODICAL. DO NOT DEVIATE.
- ❌ DO NOT try to add debug logging (crashes swallow output)
- ❌ DO NOT try to fix the issue until the exact line is found
- ❌ DO NOT skip steps or make assumptions
- ❌ DO NOT re-enable all code if something doesn't work
- ❌ DO NOT report back until the EXACT line is identified
- ✅ DO follow the bisection process EXACTLY as described
- ✅ DO continue methodically even if it seems tedious
- ✅ DO test after EVERY single change
Required Information from User
Ask the user to provide:
- Test command: The exact command that reproduces the crash (e.g.,
zig build test-opcodes -Dtest-filter='ADD opcode') - File path: The file suspected to contain the crash (or use $ARGUMENTS if provided)
- Function name (optional): Specific function to start with if known
- Crash symptoms: What happens (segfault, panic, hang, etc.)
Phase 1: Initial Verification
- Run the test command to confirm the crash reproduces
- Note the exact error message or crash behavior
- You do not need to create a backup of the file, we already use version control for that
Phase 2: Function-Level Bisection
-
Identify all functions in the suspect file
-
Replace function bodies with stubs one by one, testing after each:
pub fn someFunction(args: Type) ReturnType { // [entire function body commented out] // Return stub value to keep compilation working: return undefined; // or appropriate default/zero value }For error unions:
return error.NotImplemented;For optionals:return null;For void: just return For noreturn:unreachable;or@panic("stub"); -
When the crash disappears after stubbing a function, YOU FOUND THE CULPRIT FUNCTION
-
Restore all other functions to their original state
-
Keep only the culprit function for investigation
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 231 lines · 16 tokens per session scan A 2b5423e17d01
crash-bisection is a command published in the GitHub repository evmts/guillotine-mini (11 stars, last pushed 2mo ago), licensed MIT. It adds 16 tokens to every session and 2,013 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
crash-bisection compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| crash-bisectionevmts/guillotine-mini | Command | 11 | 2mo ago | MIT | A | 16 | Claude Code |
| qa-changesOpenHands/extensions | Command | 176 | yesterday | MIT | A | 63 | Claude Code |
| react-nativegetsentry/toolkit | Command | 957 | yesterday | none | A | 0 | Claude Code |
| auditar-cascadingluanpdd/kit-mcp | Command | 5 | 5d ago | MIT | A | 54 | Claude Code |
| http-serviceMadAppGang/magus | Command | 14 | yesterday | MIT | A | 27 | Claude Code |
| doctordoncheli/don-cheli-sdd | Command | 59 | 2mo ago | Apache-2.0 | A | 76 | Claude Code |
| doctorfatihkan/badi | Command | 7 | 1mo ago | MIT | A | 0 | Claude Code |