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 agentmods add skills/luisfurquim/wings/sec-hostile-inputnpx skills add luisfurquim/wings --skill sec-hostile-inputgit clone --depth 1 https://github.com/luisfurquim/wingsWrote 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/luisfurquim/wings/sec-hostile-input)<a href="https://agentmods.dev/skills/luisfurquim/wings/sec-hostile-input"><img src="https://agentmods.dev/badge/skills/luisfurquim/wings/sec-hostile-input.svg" alt="Measured on agentmods" 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 | $0.00062 | $0.00842 |
| Opus 5 | $0.00031 | $0.00421 |
| Sonnet 5 | $0.00012 | $0.00168 |
| Haiku 4.5 | $0.00006 | $0.00084 |
Grade A, and why
sec-hostile-input 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 6d 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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hostile Input & Parsing Discipline
Two converging sources:
- djb: parsing is where security bugs concentrate. Avoid parsing; when unavoidable, parse once, at the boundary, into a typed structure, and let interior code consume only the typed result. Prefer formats too simple to misparse.
- Margaret Hamilton (Apollo): "the operator is trained and won't make mistakes" is not a defense. Her prelaunch-program bug report was dismissed on exactly that argument — then Jim Lovell triggered it on Apollo 8 (P01 wipe). Validate everyone's input: end users, webdevs, and our own build artifacts.
Rules
- Parse at the trust boundary, never again. Each input crosses exactly
one validator that produces a typed Go value. Downstream functions take
the typed value, not the raw string. If a function deep in the call tree
does
strings.Spliton data that arrived from outside, the boundary leaked. - Reject, don't repair. Malformed input → typed
errorand stop (feedback_errors_over_panic: public API returns error; the app dev decides whether to panic). Fixing up input silently creates two parsers: the real one and the one in the attacker's head. Documented exception:expr.ParseFlexBlockleniency deliberately falls back to literal passthrough (render-as-text), which is a fail-closed default, not repair. - Prefer dumb formats. JSON via
encoding/jsoninto a concrete struct (DisallowUnknownFieldswhen the schema is closed) beats any hand-rolled format. If you design a format, make it a fixed shape, not a language. - Validate the meaning, not just the shape. BCP-47 tags go through the
guard (don't interpolate raw lang strings into URLs/paths); indices into
catalogs are bounds-checked; file paths from config are cleaned and
confined to the expected root (the dev server's dotfile/key blocking in
cmd/buildis the model). - Webdev input is input. Template attributes, binding names, wings.json,
WINGS_*env vars. Silent failure is the worst outcome — the camelCase binding lint incmd/build/lint.goexists because the HTML parser lowercases attribute names and the mismatch failed silently. New template-facing features must either work or fail the build with file:line; never render wrong quietly.
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.
- 6d ago First seen · 65 lines · 62 tokens per session scan A c6a9c1c7b04c
sec-hostile-input is a skill published in the GitHub repository luisfurquim/wings (29 stars, last pushed 1mo ago), licensed MPL-2.0. It adds 62 tokens to every session and 842 once invoked, about $0.0003 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
chronicle
Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…