Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add fakoli/fakoli-plugins/plugin install gate-routerWrote 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/fakoli/fakoli-plugins/gate-check)<a href="https://agentmods.dev/skills/fakoli/fakoli-plugins/gate-check"><img src="https://agentmods.dev/badge/skills/fakoli/fakoli-plugins/gate-check.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 | $0.00110 | $0.00844 |
| Opus 5 | $0.00055 | $0.00422 |
| Sonnet 5 | $0.00022 | $0.00169 |
| Haiku 4.5 | $0.00011 | $0.00084 |
Grade C, and why
gate-check scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
shell, so a changed file named `x;rm -rf ~` is an inert argument), but the How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Gate Check
Deterministic changed-path → verify-command routing. The rules live in the
project, not in anyone's memory: .claude/gate-router.local.md (the
marketplace's plugin-settings pattern).
Check the current diff
bash "${CLAUDE_PLUGIN_ROOT}/scripts/gate-router.sh" --list # what's required
bash "${CLAUDE_PLUGIN_ROOT}/scripts/gate-router.sh" --run # run them, stop on failure
bash "${CLAUDE_PLUGIN_ROOT}/scripts/gate-router.sh" --list --json
- Default diff base is
origin/main(fall backHEAD); override with--base <ref>. The changed set is committed-vs-base + staged + unstaged — the full "about to ship" surface. --runexecutes gates in rule order and stops at the first failure with its exit code — report that failure to the user verbatim, fix, re-run.- No config → the script says so and exits 0 (nothing to enforce). Suggest authoring one (below) when the repo clearly has gate-worthy surfaces.
Author or extend the rules file
.claude/gate-router.local.md, rules in the frontmatter, glob => command:
---
rules:
- docs/** => mkdocs build --strict
- "**/*.sh" => bash -n {files}
- bin/src/** => cd bin && uv run pytest -q
- plugins/** => bash scripts/validate.sh
---
Notes for humans below the fence (ignored by the router).
**crosses directories; a leading**/also matches paths at the root.{files}passes the matched files to the command as separate arguments (safe for names with spaces/metacharacters) — use it for linters; omit it for suite commands that don't take a file list.- Duplicate commands from overlapping rules run once. Order = rule order.
- Derive rules from the repo's own history: CI workflow steps, CLAUDE.md test instructions, and past incident classes (encoding, docs links) are the gate candidates. Keep each gate FAST — these run per-change, not nightly.
Trust boundary
The rules file executes shell commands — treat .claude/gate-router.local.md
like a Makefile: only run --run in repos you trust, and review a rules file
you didn't write before running it. The script is injection-safe with respect
to filenames (matched files are passed as argv, never interpolated into the
shell, so a changed file named x;rm -rf ~ is an inert argument), but the
COMMANDS themselves are run verbatim with your privileges.
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.
- 2d ago First seen · 74 lines · 110 tokens per session scan C 3e92fb26674c
gate-check is a skill published in the GitHub repository fakoli/fakoli-plugins (4 stars, last pushed 27d ago), licensed MIT. It adds 110 tokens to every session and 844 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
better-auth-knowledge-patch
Use this skill when implementing, upgrading, debugging, or reviewing Better Auth applications. Start with the quick references below, then load the topic file that matches the task.
clickhouse-knowledge-patch
Use this skill when writing, reviewing, upgrading, or operating ClickHouse SQL and server configurations. Start with the compatibility-sensitive items below, then load the topic reference that matches the task.
arch-knowledge-patch
Restart the daemon immediately after upgrading openssh to 9.8p1 because the old daemon cannot accept new connections.
axum-knowledge-patch
Axum 0.8 uses return-position impl Trait in FromRequestParts and FromRequest. Implement their methods with native async fn; do not retain the old #[asynctrait] annotation.
caddy-knowledge-patch
Use this patch when writing or reviewing Caddyfile or JSON configuration, upgrading Caddy, building custom binaries, or operating Caddy's HTTP, proxy, certificate, and telemetry features.
cloudflare-d1-knowledge-patch
Use this skill when designing, implementing, migrating, operating, or debugging Cloudflare D1 databases. Check the project’s Wrangler configuration, binding usage, database generation, account plan, and remote-versus-local command flags before applying guidance.