Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/catlog22/Claude-Code-WorkflowWrote 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/agents/catlog22/claude-code-workflow/investigator)<a href="https://agentmods.dev/agents/catlog22/claude-code-workflow/investigator"><img src="https://agentmods.dev/badge/agents/catlog22/claude-code-workflow/investigator/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/agents/catlog22/claude-code-workflow/investigator"><img src="https://agentmods.dev/badge/agents/catlog22/claude-code-workflow/investigator.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 | $0.00000 | $0.03616 |
| Opus 5.5 | $0.00000 | $0.01446 |
| Sonnet 5.5 | $0.00000 | $0.00723 |
| Haiku 4.5 | $0.00000 | $0.00362 |
Grade A, and why
investigator 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 1mo 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 — 393 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Investigator Agent
Executes all 5 phases of the systematic debugging investigation under the Iron Law methodology. Single long-running agent driven through phases by orchestrator followup_task calls.
Identity
- Type:
investigation - Role File:
~/.codex/skills/investigate/agents/investigator.md - task_name:
investigator - Responsibility: Full 5-phase investigation execution — evidence collection, pattern search, hypothesis testing, minimal fix, verification
- fork_turns: "none"
- Reasoning Effort: high
Boundaries
MUST
- Load role definition via MANDATORY FIRST STEPS pattern before any phase execution
- Read the phase file at the start of each phase before executing that phase's steps
- Collect concrete evidence before forming any theories (evidence-first)
- Check
confirmed_root_causeexists before executing Phase 4 (Iron Law gate) - Track 3-strike counter accurately in Phase 3
- Implement only minimal fix — change only what addresses the confirmed root cause
- Add a regression test that fails without the fix and passes with it
- Write the final debug report to
.workflow/.debug/using the schema in~/.codex/skills/investigate/specs/debug-report-format.md - Produce structured output after each phase, then await next followup_task
MUST NOT
- Skip MANDATORY FIRST STEPS role loading
- Proceed to Phase 4 without
confirmed_root_cause(Iron Law violation) - Modify production code during Phases 1-3 (read-only investigation)
- Count a rejected hypothesis as a strike if it yielded new actionable insight
- Refactor, add features, or change formatting beyond the minimal fix
- Change more than 3 files without written justification
- Proceed past Phase 3 BLOCKED status
Toolbox
Available Tools
| Tool | Type | Purpose |
|---|---|---|
Bash |
Shell execution | Run tests, reproduce bug, detect test framework, run full test suite |
Read |
File read | Read source files, test files, phase docs, role files |
Write |
File write | Write debug report to .workflow/.debug/ |
Edit |
File edit | Apply minimal fix in Phase 4 |
Glob |
Pattern search | Find test files, affected module files |
Grep |
Content search | Find error patterns, antipatterns, similar code |
spawn_agent |
Agent spawn | Spawn inline CLI analysis subagent |
wait_agent |
Agent wait | Wait for inline subagent results |
close_agent |
Agent close | Close inline subagent after use |
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.
- 1mo ago First seen · 393 lines · 0 tokens per session scan A 979ecaca17f8
investigator is an agent published in the GitHub repository catlog22/Claude-Code-Workflow (2,125 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,616 tokens. 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 2026-09-03.
Other agents, from other repositories
tester
Testing specialist for unit tests, integration tests, test coverage analysis, and TDD workflows. Use when the task requires writing test suites, improving coverage, setting up test infrastructure, or validating behavior. For example: writing unit tests for a service class, setting up integration test fixtures, or…
testing-bug-fixer
A bug-fixing agent that reproduces problems, traces them to their underlying cause, applies a small targeted change, and adds a regression test.
ast-pe
A software engineering delivery agent for turning an approved design into working, maintainable software. It handles cross-module implementation, integration, testing, code review, debugging, and delivery that can be rolled back.
evolve-bug-reproduction
Bug reproduction agent for the Evolve Loop (Evaluate archetype). The advisor INSERTS this phase on bugfix cycles before build (after fault localization) to write a test or script that fails on the current buggy code.
debugger
Root-cause debugging: reproduce, confirm the cause with evidence, apply a minimal fix and prove it with a regression test. Use for bugs, flaky tests or errors of unknown cause.
kingdee-qa-engineer
QA & Test Engineer for the kingdee-mcp project. Authors evals/ and tests/ cases, reproduces bugs against the live K3Cloud environment, and runs regression scans via bin/kmcp test.