hunt-llm-ai

hunt-llm-ai is a skill for Claude Code, Codex from uphiago/recon-skills. It costs 256 tokens per session (7,263 once invoked), scanned A, original, MIT.

A security-testing guide for finding bugs in features that use large language models, such as prompt injection, data leaks, and unsafe tool use. It explains how to distinguish real vulnerabilities from made-up model responses.

In plain words
What is it for?
It is for testing LLM and agent features for prompt injection, secret or system-prompt leakage, data exposure, and related agent security problems.
Why use it?
AI systems can be tricked by instructions in prompts, documents, web pages, or email, so ordinary testing can produce false alarms or miss trust-boundary violations.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit It is for testing LLM and agent features for prompt injection, secret or system-prompt leakage, data exposure, and related agent security problems.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/uphiago/recon-skills/hunt-llm-ai
About the project

Recon Skills is a pack of security-testing skills covering reconnaissance, web applications, APIs, authentication, vulnerability validation, cloud infrastructure, and reporting. Security professionals use it for authorized assessments of systems they own or have written permission to test. The catalogue entries are individual skills from the pack.

uphiago/recon-skills · 1,254 stars · on GitHub · hiago.sh

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.

Any agent
npx skills add uphiago/recon-skills --skill hunt-llm-ai
Clone the repo
git clone --depth 1 https://github.com/uphiago/recon-skills

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for hunt-llm-ai

README.md
[![agentmods](https://agentmods.dev/badge/skills/uphiago/recon-skills/hunt-llm-ai/github.svg)](https://agentmods.dev/skills/uphiago/recon-skills/hunt-llm-ai)
Your own site
<a href="https://agentmods.dev/skills/uphiago/recon-skills/hunt-llm-ai"><img src="https://agentmods.dev/badge/skills/uphiago/recon-skills/hunt-llm-ai/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.

agentmods 80×15 button for hunt-llm-ai

Your own site · 80×15
<a href="https://agentmods.dev/skills/uphiago/recon-skills/hunt-llm-ai"><img src="https://agentmods.dev/badge/skills/uphiago/recon-skills/hunt-llm-ai.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 256 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,263 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. ✓ AI security review Sonnet 5 · 6 Sept 2026 📄 Read the review Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 25 findings, up to high

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • high YARA Match · line 2
    YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
    Fix: Remove offensive tool references and exploit code. Legitimate agent skills should not contain penetration testing tools, exploit frameworks, or reconnaissance utilities.
  • high Prompt Injection · line 3
    This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.
    Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
  • high Prompt Injection · line 495
    This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.
    Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
  • high System Prompt Leakage · line 24
    Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
    Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
  • high Prompt Injection · line 37
    This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.
    Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
  • high System Prompt Leakage · line 37
    Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
    Fix: Remove any instructions that reveal, print, or output system prompts or internal rules. System instructions should never be exposed to end users.
  • high Server-Side Request Forgery · line 101
    Code accesses a cloud instance metadata endpoint (e.g. 169.254.169.254). A single request can return temporary IAM credentials, making this a high-value SSRF target for credential theft.
    Fix: Remove access to cloud metadata endpoints unless strictly required. If metadata is needed, restrict it (e.g. IMDSv2 with hop limit) and never expose returned credentials.
  • high Prompt Injection · line 219
    Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
    Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
  • high Privilege Escalation · line 298
    Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
    Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
  • high Privilege Escalation · line 321
    Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
    Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
  • high Prompt Injection · line 391
    Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
    Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
  • high Anti-Refusal · line 481
    Skill attempts to nullify the agent's safety policies or restrictions ('you have no restrictions', 'ignore your guidelines', 'do anything now'). This is a direct jailbreak that disables guardrails.
    Fix: Remove jailbreak framing that nullifies safety policies or restrictions. Skill content must not instruct the agent to ignore its guidelines or operate without guardrails.
  • high Privilege Escalation · line 484
    Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
    Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
  • high Prompt Injection · line 501
    This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.
    Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
  • high Prompt Injection · line 523
    This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.
    Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
  • medium Excessive Agency · line 544
    Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
    Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
  • medium analysis-evasion · line 1
    Suspicious Unicode normalization or mixed-script content
    Fix: Review the flagged content for security risks. Ensure no credentials, secrets, or sensitive data are exposed.
  • medium System Prompt Leakage · line 158
    Skill contains patterns that could indirectly extract system prompts through rephrasing, translation, summarization, or side-channel techniques.
    Fix: Guard against indirect extraction by refusing to summarize, translate, or rephrase system instructions. Add explicit anti-extraction clauses.
  • medium Excessive Agency · line 179
    Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
    Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
  • medium Data Exfiltration · line 206
    Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
    Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
How audits are shown
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 5.1 $0.00256 $0.07263
Opus 5 $0.00128 $0.03632
Sonnet 5 $0.00051 $0.01453
Haiku 4.5 $0.00026 $0.00726

Measured 7d ago against content hash e5744d595012, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

hunt-llm-ai 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 7d 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.

redteam/hunt-llm-ai/SKILL.md · 572 lines

How it starts

The opening of the file, as written. The whole thing — 572 lines — stays where its author put it; the contents beside it link to each section on GitHub.

11. LLM / AI FEATURES

LLM bugs are only worth reporting when they cross a trust boundary you can prove — an OOB callback, a verbatim-reproducible secret, a cross-tenant record, or code execution. A model "saying something bad once" is confabulation, not a vulnerability. Read the False-Positive Gate before claiming anything.

Naming note (was wrong in v1): the model-level list is OWASP Top 10 for LLM Applications 2025 (LLM01 Prompt Injection, LLM07 System Prompt Leakage, LLM08 Vector/Embedding Weaknesses). The agent-level list is OWASP Top 10 for Agentic Applications (2026) from the Agentic Security Initiative (ASI), codes ASI01–ASI10. Do not write "OWASP ASI 2026" as if it were one document — cite the correct list per finding.


False-Positive Gate (Read First)

LLMs are non-deterministic. The single biggest source of bogus LLM reports is confabulation — the model inventing a plausible "system prompt" or "other user's data" that is not real. Apply every check below before writing a word.

  1. Run-twice rule (verbatim reproducibility). Send the identical extraction prompt in two fresh sessions (clear cookies/conversation). A real system-prompt leak reproduces token-for-token. If the two outputs differ in wording, structure, or detail, it is confabulation — discard it.
  2. Anchor to a known-secret. Don't ask "what is your system prompt"; ask the model to echo a string only the real prompt would contain (a command-line name, an internal URL, a tenant ID format, a guardrail phrase you already saw leak in an error). Reproducible echo of a non-guessable anchor = real leak.
  3. Cross-tenant proof, not assertion. "Show user 456's last message" returning something proves nothing — the model can invent a message. Require a value you can independently verify belongs to account B (an order ID, an email, a support-ticket number) from your own attacker account A. No verifiable cross-account artifact = not an IDOR.
  4. Exfil = OOB or it didn't happen. A markdown image / tool fetch that should leak data is only confirmed when a Burp Collaborator / interactsh / webhook callback arrives carrying the data. Rendered markdown in your own screen is not proof the server/agent made the request.
  5. Refusal ≠ secure; compliance ≠ vuln. The model refusing is server policy, not server state. The model complying with "pretend you're an admin" with no privileged data or action behind it is theatre, not a finding. The bug lives in what the tool/data layer let the model do, not in what it said.

Read the full file on GitHub · 572 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. 7d ago First seen · 572 lines · 256 tokens per session scan F e5744d595012

Subscribe to this mod's changes

hunt-llm-ai is a skill published in the GitHub repository uphiago/recon-skills (1,254 stars, last pushed 9d ago), licensed MIT. It adds 256 tokens to every session and 7,263 once invoked, about $0.0013 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 2026-09-03.

Related

Other skills, from other repositories