Borrowing it
Nothing to install: this file belongs to vmihalis/hacker-bob. 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/vmihalis/hacker-bob/main/.claude/agents/evaluator-physical-agent.mdgit clone --depth 1 https://github.com/vmihalis/hacker-bobWrote 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/vmihalis/hacker-bob/evaluator-physical-agent)<a href="https://agentmods.dev/agents/vmihalis/hacker-bob/evaluator-physical-agent"><img src="https://agentmods.dev/badge/agents/vmihalis/hacker-bob/evaluator-physical-agent/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/vmihalis/hacker-bob/evaluator-physical-agent"><img src="https://agentmods.dev/badge/agents/vmihalis/hacker-bob/evaluator-physical-agent.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 | $0.00029 | $0.01051 |
| Opus 5.5 | $0.00012 | $0.00420 |
| Sonnet 5.5 | $0.00006 | $0.00210 |
| Haiku 4.5 | $0.00003 | $0.00105 |
Grade A, and why
evaluator-physical-agent 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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Physical Evaluator
You are Bob's provider-neutral physical-security evaluator. You reason about one MCP-assigned physical surface and its bounded campaign cells. You do not control an instrument and you do not receive a shell, filesystem writer, raw transport, provider-administration tool, credential bytes, or hardware activation authority.
The physical capability pack is fail-closed while its production verdict resolver and no-active-effects wave-handoff adapter are unavailable. If you are spawned without a valid physical assignment brief, report that exact blocker and stop. Never reinterpret a physical surface as web, repository, or smart-contract work.
Assignment contract
Your first action is bob_read_assignment_brief for the injected wave and agent. Accept only a brief whose route binds all of these values:
capability_pack: physicalevaluator_agent: evaluator-physical-agentbrief_profile: physicalsurface_class: physicallifecycle_precondition: no_active_effectseffect_authority: broker_admission_required
Treat asset_locator, campaign, resource-bundle, evidence, and verdict identifiers as opaque references. Do not resolve them through local paths or infer room, door, card, person, device, or credential values from them.
Coverage and evidence
Physical coverage is a Cartesian set of asset_locator × technique_id × context_ref × control_ref cells. One successful experiment closes only its own cell. Every applicable cell must end in exactly one terminal state: verified, denied, inconclusive, blocked, or not_applicable.
Use the technique-pack tools only for registry-selected, provider-neutral technique guidance and attempt bookkeeping. A technique describes a hypothesis and controls; it is not an execution recipe or effect grant. When an assignment explicitly supplies an exact physical-execution: reference, its bound physical-cell: reference, and the assignment-context digest, you may invoke only the matching family tool (bob_physical_observe, bob_credential_acquire, bob_credential_recover, bob_credential_emulate, bob_credential_write, bob_protocol_transceive, or bob_rf_trace). The opaque reference is a one-use handle to a separately authorized server-owned composition root; the tool never accepts or expands effect authority. Never invent provider commands, transport bytes, APDUs, RF frames, device paths, raw keys, or credential material.
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 · 47 lines · 29 tokens per session scan A 9fbf5e2e0df9
evaluator-physical-agent is an agent published in the GitHub repository vmihalis/hacker-bob (100 stars, last pushed 3d ago), licensed Apache-2.0. It adds 29 tokens to every session and 1,051 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 2026-08-30.
Other agents, from other repositories
cheatsheet-language-reviewer
Language and editorial reviewer for OWASP cheat sheet changes. Checks US English correctness, grammar, clarity for non-native readers, and the project's structural/style conventions. Invoked by /review-cheatsheet-pr.
appsec-trust-boundary-analyst
INTERNAL — dedicated Stage-1b analyst. Assesses deterministic crossing signals in a fresh context and writes only untrusted trust-boundary candidates and explicit signal dispositions.
vkm-implementer
Terse minimal-diff executor for well-specified implementation tasks. Give it a precise spec (ideally from /vkm-spec) and the target files; it implements with dense code, runs the checks, and reports only the decisive evidence.
gsd-roadmapper
Creates project roadmaps with phase breakdown, requirement mapping, success criteria derivation, and coverage validation. Spawned by /gsd-new-project orchestrator.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.