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.
npx skills add UnboundCompute/security-agent-skills --skill hunting-search-engine-injectiongit clone --depth 1 https://github.com/UnboundCompute/security-agent-skillsWrote 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/skills/unboundcompute/security-agent-skills/hunting-search-engine-injection)<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/hunting-search-engine-injection"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/hunting-search-engine-injection/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/skills/unboundcompute/security-agent-skills/hunting-search-engine-injection"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/hunting-search-engine-injection.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 | $0.00166 | $0.02034 |
| Opus 5 | $0.00083 | $0.01017 |
| Sonnet 5 | $0.00033 | $0.00407 |
| Haiku 4.5 | $0.00017 | $0.00203 |
Grade A, and why
hunting-search-engine-injection 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.
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.
Hunting search-engine injection: when a search query is attacker-shaped
Search and analytics engines take rich structured queries, and applications frequently build those queries
from user input. The injection surface has three shapes. First, the query DSL body: when a request object
is merged into the query JSON, the attacker can add filter clauses, select fields the caller was not meant
to read, or attach aggregations that reveal data across the index. Second, the raw query string: a
query_string or Lucene query parameter exposes operators, field selectors, and wildcards, so a caller who
controls it can query fields and ranges outside the intended scope. Third, scripting: inline or stored
scripts run inside the engine, and untrusted input reaching a script is code execution in the cluster. You
find these by separating bound search terms from query structure and scripts, and tracing untrusted input
into the latter.
When to use
- An application forwards user input into a search or analytics cluster as query JSON or a query string.
- A request object is merged into the query DSL, or a raw Lucene/query-string parameter is caller-controlled.
- Inline or stored scripts run in the engine and untrusted input can reach a script or its parameters.
Scope check
Test search-engine injection only against clusters and applications you own or are authorized to assess, on non-production data. A confirming query or script can read across indices or run code in the cluster, so stay inside the authorized boundary. If you can't name the authorization, stop.
The loop
- Establish the query and script sinks and split them. Inventory every search call and separate the parts that bind a search term from the parts that carry structure: the query DSL body, field and index selectors, aggregations, raw query-string parameters, and inline or stored scripts. This is the false-positive killer: a query that binds the user's term into a fixed match clause over a fixed field is not injectable. Name the structure-carrying and script sinks first.
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.
- 7d ago First seen · 137 lines · 166 tokens per session scan A fd7b32a8381c
hunting-search-engine-injection is a skill published in the GitHub repository UnboundCompute/security-agent-skills (5 stars, last pushed 3d ago), licensed MIT. It adds 166 tokens to every session and 2,034 once invoked, about $0.0008 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 2026-09-05.
Other skills, from other repositories
prodcheck-review
Review this codebase against the prodcheck pre-production checklists — security, performance, scale, integrations and post-launch readiness. Use when asked to check whether a project is ready to ship, to audit an area before launch, or to work through a specific checklist. Produces evidence with file:line citations…
integrate-arcjet-guard-genkit
Integrate Arcjet security into a Genkit JS agent using @arcjet/guard — wrap ai.defineTool, put guardMiddleware on generate({ use }) for unwrapped / MCP / filesystem tools, and read a caller-owned id from generate({ context }). Use when asked to add Arcjet to genkit, rate limit its tools, screen inbound messages, or…
integrate-arcjet-guard-claude-managed-agents
Integrate Arcjet security into Claude Managed Agents (hosted REST+SSE, beta managed-agents-2026-04-01) using @arcjet/guard — screen user.message / initialevents before sessions.events.send, and gate custom tools on agent.customtooluse. Use when asked to add Arcjet to Claude Managed Agents, rate limit custom tools, or…
linear-claude-skill
Manage Linear issues, projects, and teams.
panguard
AI agent security platform — audit skills, scan for threats, and run 24/7 protection with 9,700+ detection rules.
security-guide
A security deployment guide for OpenClaw, with instructions in English and Chinese. It covers network exposure, container isolation, credentials, permissions, and audit logging.