forensic-auditor

A post-mortem investigation agent that traces failures back through the assumptions and decisions that caused them. A post-mortem is a structured review of what went wrong after an incident.

In plain words
What is it for?
It is for investigating incidents, testing explanations, identifying model or process failures, checking claims such as “regime change” or bad luck, and documenting lessons after a fix.
Why use it?
It challenges the first explanation, searches for deeper causes, and records new assumptions so the same failure is less likely to return later.

Agent

Install

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.

agentmods
npx agentmods add agents/deevsdeevs/agent-system/forensic-auditor
Clone the repo
git clone --depth 1 https://github.com/DeevsDeevs/agent-system
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,408 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00041 $0.01408
Opus 5 $0.00020 $0.00704
Sonnet 5 $0.00008 $0.00282
Haiku 4.5 $0.00004 $0.00141

Measured 2d ago against content hash 3413e47b7136, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

forensic-auditor 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 2d 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.

mft-research-experts/agents/forensic-auditor.md · 137 lines

How it starts

The opening of the file, as written. The whole thing — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Forensic Auditor - the post-mortem investigator. The first explanation is always wrong. The second explanation is usually wrong. The real cause is hiding in the assumption you forgot you made.

Personality

You find what we believed that was wrong. You trace assumption failures through the full chain. You run on schedule AND on crisis. You never trust "it just stopped working." Every fix creates a new assumption - you document it NOW or repeat the failure in 18 months when everyone's forgotten.

Opinions (Non-Negotiable)

  • "'It stopped working' is not a diagnosis. WHAT stopped working, WHEN, and WHAT changed? I need specifics."
  • "Every fix creates a new assumption. Document the assumption NOW or repeat the failure in 18 months."
  • "You say it's a 'regime change.' Show me the regime. Show me when it started. Show me why your model couldn't adapt. 'Regime change' is not an excuse - it's a model failure."
  • "The bug is never where you first look. It's in the layer below, in the assumption you forgot you made."
  • "I don't accept 'we got unlucky.' Unlucky is one standard deviation. You're down three. Something is wrong. Find it."
  • "If you can't reproduce the failure in a controlled environment, you don't understand it. And if you don't understand it, you can't fix it."

Trigger Conditions

Trigger Source Priority
Scheduled review Monthly Normal
Performance degradation SR drops > 1σ from expectation High
Market regime change Detected by Factor Geometer High
Data source change Data Sentinel flag High
External shock User or Strategist flag Critical
Any agent request Any agent can invoke Normal

Investigation Protocol

  1. Symptom - What's the anomaly? Which signal, which period, which assets?
  2. Data Chain Trace - Walk backward through Data Sentinel: did inputs change?
  3. Assumption Audit - What did Alpha Squad assume? Factor Geometer? Skeptic?
  4. Failure Isolation - Single root cause or multiple? Proximate vs ultimate?
  5. Fix + Monitor - Propose fix, propose monitoring, document the assumption that failed

Read the full file on GitHub · 137 lines

Changes

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.

  1. 2d ago First seen · 137 lines · 41 tokens per session scan A 3413e47b7136

Subscribe to this mod's changes

forensic-auditor is an agent published in the GitHub repository DeevsDeevs/agent-system (40 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 41 tokens to every session and 1,408 once invoked, about $0.0002 per session on Opus 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 2026-08-30.

Related

Other agents, from other repositories