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 ShockRock2004/kamar-taj --skill dormammugit clone --depth 1 https://github.com/ShockRock2004/kamar-tajWrote 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/shockrock2004/kamar-taj/dormammu)<a href="https://agentmods.dev/skills/shockrock2004/kamar-taj/dormammu"><img src="https://agentmods.dev/badge/skills/shockrock2004/kamar-taj/dormammu/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/shockrock2004/kamar-taj/dormammu"><img src="https://agentmods.dev/badge/skills/shockrock2004/kamar-taj/dormammu.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 | $0.00140 | $0.02190 |
| Opus 5 | $0.00070 | $0.01095 |
| Sonnet 5 | $0.00028 | $0.00438 |
| Haiku 4.5 | $0.00014 | $0.00219 |
Grade B, and why
dormammu scanned grade B 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 9d 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.
Subtle steeringmediumPrompt injection
Instructions that bias recommendations or shape behaviour without the user noticing.
clear the goal yourself to escape the loop, and you never tell the user to clear a How it starts
The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dormammu Protocol — bargain until it is actually correct
IRON RULE — read this first
When /dormammu is invoked, this protocol is an IRON RULE for the task and must
be followed at all costs, no matter what. You do not get to skip the loop, cut it
short, or declare the work done to save time or effort. No premature "done". No
unverified claims. No "this probably works." You keep bargaining — implement,
attack your own work, revert, retry — until the result is proven correct by
actually running the verification. Quitting early, hand-waving, or calling a first
draft final is forbidden.
This binds how rigorously and relentlessly you work. It strengthens correctness, safety, and honesty — it never licenses overriding them. The only honest way to stop short is the guardrails below: if you are genuinely blocked by something only the user can resolve, or the request itself is flawed, you say so plainly. You never fake completion.
This is meant to be backed by /goal, but you do not run /goal yourself.
/goal is a user-only UI command — the Skill tool cannot invoke it, and trying
to will just error. So in Phase 1 you compose the exact /goal command and hand
it to the user to paste. If they paste it, /goal arms a session Stop hook: a wall
the harness puts up that refuses to let the turn end until the condition holds, and
clears the moment it does — "do not stop early" stops being willpower and becomes a
wall. If they do not paste it, you still follow this protocol to the letter; the
rule binds you either way. Either way, never call /goal via the Skill tool.
Dormammu and /goal are the same idea from two directions: /goal can make the
loop impossible to escape, and this protocol makes the loop productive whether or
not the wall is up.
A disciplined loop for serious work. The point is to refuse premature "done": keep auditing and reverting until the solution genuinely survives scrutiny and verification, not just until a first draft compiles or a checklist is ticked.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 9d ago First seen · 160 lines · 140 tokens per session scan B c3d7f865bfa0
dormammu is a skill published in the GitHub repository ShockRock2004/kamar-taj (4 stars, last pushed 1mo ago), licensed MIT. It adds 140 tokens to every session and 2,190 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it B with 1 finding (subtle steering). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
senior-engineering-partner
A strict code reviewer, pair programmer, debugger, and mentor for Python, Bash, Google Apps Script, JavaScript, and Swift/Apple platforms. Use when writing, reviewing, debugging, planning, or securing code, or for senior-level rigor, a security review, or mentoring. Mode triggers — REVIEW: (critique + refactor)…
yolo-debug
Use when a reported problem, failure, or bug needs a systematic root-cause investigation — reproduce → isolate → fix → verify — recorded as one durable artifact. Triggers on "fix the 500 on login", "X is broken", "why is Y failing?", "debug this", "track down this bug".
performance-optimization
Web and backend performance optimization including Core Web Vitals (LCP, FID, INP, CLS), bundle analysis with webpack-bundle-analyzer, code splitting, lazy loading, image optimization (WebP, AVIF, responsive srcset), caching strategies (HTTP Cache-Control, CDN, Redis, application-level), database query optimization…
go-linter-configuration
Configure and troubleshoot golangci-lint for Go projects. Includes complete .golangci.yml examples, import resolution fixes, CI optimization, and linter selection workflows.
refactoring-patterns
Code refactoring patterns with before/after diffs for extract method, extract class, inline, move, rename, dead code removal, dependency injection, and code smell detection.
golang-safety
Defensive Golang coding to prevent panics, silent data corruption, and subtle runtime bugs. Use whenever writing or reviewing Go code that involves nil-prone types (pointers, interfaces, maps, slices, channels), numeric conversions, resource lifecycle (defer in loops), or defensive copying. Also triggers on questions…