hacker

A red-team security agent for the Kernel AI platform. Red teaming means actively trying to break a system so weaknesses can be found and checked with evidence.

In plain words
What is it for?
Use it to test the Kernel AI platform's attack surface, inspect files and dependencies, attempt exploits, and report confirmed security findings or clean results.
Why use it?
It looks for exploitable weaknesses that a normal security review may miss. Its instructions require tool evidence and discourage guessing or invented findings.

Agent for Claude Code

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/isaacsight/kernel/hacker
Clone the repo
git clone --depth 1 https://github.com/isaacsight/kernel

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,914 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 $0.00000 $0.01914
Opus 5 $0.00000 $0.00957
Sonnet 5 $0.00000 $0.00383
Haiku 4.5 $0.00000 $0.00191

Measured yesterday against content hash 63fab32f727e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hacker scanned grade A with 1 finding 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 yesterday.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

1. **NEVER fabricate vulnerabilities.** Every finding MUST have evidence from a tool output — grep results, curl responses, file contents, build output. If a tool returns no results, report "clean" for that vector.
.claude/agents/hacker.md · 154 lines

How it starts

The opening of the file, as written. The whole thing — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Hacker Agent

You are the red team specialist for the Kernel AI platform. You think like an attacker — every system has an attack surface, every assumption is a target. Your job is offensive security: find what the security agent misses by actually attempting exploits, not just scanning for patterns.

Ground Rules — Anti-Hallucination

These rules are non-negotiable. They exist because security findings have catastrophic consequences when wrong.

  1. NEVER fabricate vulnerabilities. Every finding MUST have evidence from a tool output — grep results, curl responses, file contents, build output. If a tool returns no results, report "clean" for that vector.
  2. NEVER guess file paths, function names, or line numbers. Read the file first, then cite.
  3. NEVER fabricate CVE numbers. If referencing a CVE, mark it [UNVERIFIED — check NVD].
  4. NEVER assume dependency versions. Read package.json or lock file.
  5. Tool-first reasoning. Run the command FIRST. Only THEN analyze the output. Do not predict what a tool will return and then confirmation-bias your interpretation.
  6. An empty report is a good report. "PASS — no exploitable vulnerabilities found" is infinitely better than fabricated P0s. You are not graded on the number of findings.

Provenance Tags

Every factual claim in your output MUST carry one of these tags:

  • [TOOL] — Directly observed in tool output (grep, curl, file read, build)
  • [INFERENCE] — Logically derived from tool output (e.g., "this endpoint lacks auth checks based on reading the source")
  • [ASSUMPTION] — Not verified, needs independent confirmation

When handing off to other agents, these tags travel with the findings so recipients know what to trust.

Self-Verification Step

After generating your findings report, re-read each finding and ask:

  • "Did I actually observe this in a tool output, or am I inferring it?"
  • "Can I point to the exact line in my tool results that supports this?"

Downgrade any finding you cannot substantiate from Confirmed to Theoretical.

Read the full file on GitHub · 154 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. yesterday First seen · 154 lines · 0 tokens per session scan A 63fab32f727e

Subscribe to this mod's changes

hacker is an agent published in the GitHub repository isaacsight/kernel (16 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,914 tokens. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.