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 alexmmatos/arthur-mcp --skill add-permissiongit clone --depth 1 https://github.com/alexmmatos/arthur-mcpWrote 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/alexmmatos/arthur-mcp/add-permission)<a href="https://agentmods.dev/skills/alexmmatos/arthur-mcp/add-permission"><img src="https://agentmods.dev/badge/skills/alexmmatos/arthur-mcp/add-permission/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/alexmmatos/arthur-mcp/add-permission"><img src="https://agentmods.dev/badge/skills/alexmmatos/arthur-mcp/add-permission.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 | $0.00052 | $0.00449 |
| Opus 5 | $0.00026 | $0.00225 |
| Sonnet 5 | $0.00010 | $0.00090 |
| Haiku 4.5 | $0.00005 | $0.00045 |
Grade A, and why
add-permission 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 8d 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.
What it actually says
Wire a permission end-to-end for the feature being implemented. Full rules: .claude/rules/permissions.md and docs/DESIGN_PATTERNS.md (Guard and Decorator Authorization / Auth and Permission Context sections).
- Decide: check whether an existing permission already covers this domain and risk. If yes, reuse it and stop here — do not add a near-duplicate. If not, name the new permission
<domain>_<verb>matching the action's risk (not just its domain). - Backend, in the same change:
- Add the key to the
RolePermissionscontract. - Add it to built-in role presets in
api/src/roles/permissions.ts. - Guard the controller route with
RequirePermission(or the relevant guard).
- Add the key to the
- Frontend, in the same change:
- Add the key to the
Permissionenum andUserPermissions. - Add it to role fallback presets in
src/context/permissionPresets.ts. - Gate the nav item / route / button with
can(Permission.X). - Add a restricted empty state or disabled control for view-but-not-act cases; hide navigation only when the user cannot view the surface at all.
- Add the key to the
- Tests: cover the new guard/decorator behavior and the UI gate.
- Docs: record the decision per
docs/DESIGN_PATTERNS.md's change checklist; updatedocs/FLOWS.mdif the user journey changes.
Do not ship a frontend-only permission without backend enforcement, and do not ship backend-only protected behavior without a frontend permission model + restricted UI state. If backend support must be deferred, record the gap in docs/ROADMAP.md and keep the UI inaccessible until the backend permission exists.
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.
- 8d ago First seen · 23 lines · 52 tokens per session scan A 4416b46244e8
add-permission is a skill published in the GitHub repository alexmmatos/arthur-mcp (2 stars, last pushed 1mo ago), licensed MIT. It adds 52 tokens to every session and 449 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 2026-09-03.
Other skills, from other repositories
unreal-world-builder
Takes a validated Universal Assembly Graph (UAG) produced by the game-designer skill and realizes it in UEFN by calling this repo's own MCP tools (runtoolbelttool, qfoldit tools). Use this skill whenever a UAG needs to be turned into actual placed actors/lights/audio/particles in a live UEFN project, or when the user…
game-designer
Produces a Universal Assembly Graph (UAG) -- qFoldIT's engine-neutral JSON scene/interaction plan -- from a natural-language description of a scene, an interaction concept, or the output of another qFoldIT science skill (e.g. plant-growth, mining). Use this skill whenever the user wants to design a game world…
distributed sql
Deprecated compatibility redirect for Aurora DSQL guidance. Use when a request concerns DSQL, Aurora DSQL, distributed SQL, DSQL schemas, migrations, queries, authentication, performance, or application development.
error-handling
Read the error codes these tools raise and decide whether to retry, fix the call, or stop.
skill-authoring-workflow
Design, create, or update Seizu skillsets, skills, toolsets, and Cypher-backed tools, including the self-referential skillauthoring skillset itself.
cve-severity-analysis
Compute a CVSS score distribution and severity breakdown for CVEs published in the graph over a configurable lookback window. Uses the sandbox to run Python against graph data for percentile analysis and score statistics.