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 adjudicating-taint-pathsgit 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/adjudicating-taint-paths)<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/adjudicating-taint-paths"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/adjudicating-taint-paths/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/adjudicating-taint-paths"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/adjudicating-taint-paths.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 | $0.00128 | $0.01383 |
| Opus 5 | $0.00064 | $0.00691 |
| Sonnet 5 | $0.00026 | $0.00277 |
| Haiku 4.5 | $0.00013 | $0.00138 |
Grade A, and why
adjudicating-taint-paths 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 11d 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 — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Adjudicating taint paths: lead → decided finding
A lead is a fact about structure - "an input-shaped value can reach a dangerous sink." It is never a verdict. Adjudication is the disciplined work of deciding whether that structural possibility is a real, reachable bug on the current source, and recording the decision so it isn't re-litigated next pass.
When to use
- A scanner or candidate list flagged a sink and you must confirm or kill it.
- You spotted a sink by hand and want to know if attacker input reaches it.
- You need to kill a plausible-looking lead with evidence, not vibes.
Scope check
Authorized source only (your own, OSS, CTF, in-scope engagement). If you can't name the authorization, stop.
The loop
-
Name source and sink precisely. Which exact argument of which sink is dangerous, and what is the actual untrusted entry - a request param, header, filename, env var, deserialized field? Vague framing ("user input reaches it somewhere") is how false positives survive.
-
Trace the reverse cone into the sink. What values can flow into this sink argument? This enumerates every origin. If none trace back to an untrusted source, the lead is dead - kill it, record why.
-
Trace the forward cone from the source. Where does the untrusted value go? If it never touches the sink, the lead is dead. Forward and reverse must agree; if they don't, you mis-specified an endpoint - fix it and redo.
-
Get a witness path. The strongest evidence is a concrete
source → … → sinkpath. Good tooling returns either a witness or an honest negative ("no path"). A witness is a hypothesis to verify, not a proof. -
Confirm every hop against live source. Read the actual body of each function on the path at the commit you're adjudicating. Verify the value is genuinely carried hop-to-hop and is not: reassigned to a constant/trusted value; validated, sanitized, or encoded by a guard on the path; narrowed to a safe type or bounded before the sink; or never actually passed by any caller.
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.
- 11d ago First seen · 113 lines · 128 tokens per session scan A 8e6143228d3d
adjudicating-taint-paths is a skill published in the GitHub repository UnboundCompute/security-agent-skills (5 stars, last pushed 2d ago), licensed MIT. It adds 128 tokens to every session and 1,383 once invoked, about $0.0006 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 2026-08-31.
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-langgraph
Integrate Arcjet security into a LangGraph Graph API agent using @arcjet/guard — wrap tool() / StructuredTool, wrap ToolNode for unwrapped MCP tools, and read threadid for correlation. Use when asked to add Arcjet to a LangGraph StateGraph / ToolNode agent, rate limit its tools, screen inbound messages, or block…
integrate-arcjet-guard-tanstack-ai
Integrate Arcjet security into a TanStack AI chat() app using @arcjet/guard — put guardMiddleware first on chat({ middleware }) so onBeforeToolCall gates tools, and read a caller-owned id from chat({ context }). Use when asked to add Arcjet to TanStack AI, rate limit its tools, screen inbound messages, or block prompt…
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.