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 kimmingul/xllm/plugin install xllmWrote 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/kimmingul/xllm/review)<a href="https://agentmods.dev/skills/kimmingul/xllm/review"><img src="https://agentmods.dev/badge/skills/kimmingul/xllm/review.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 | $0.00102 | $0.00838 |
| Opus 5 | $0.00051 | $0.00419 |
| Sonnet 5 | $0.00020 | $0.00168 |
| Haiku 4.5 | $0.00010 | $0.00084 |
Grade A, and why
review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
review — cross-vendor deliberation (xllm)
One entry point, four modes. Your host CLI synthesizes; external CLIs advise.
Resolve the script
Claude Code: "${CLAUDE_PLUGIN_ROOT}/scripts/xllm-review.js".
Other hosts: <plugin-root>/scripts/xllm-review.js (two dirs above this SKILL.md).
Modes
node <xllm-review.js> roles p1,p2[,p3] "<prompt>" # parallel advisors, host synthesis — NOT measured
node <xllm-review.js> blind p1,p2[,p3] "<question>" [--tiebreak] [--ready=a,b,c]
node <xllm-review.js> debate p1,p2[,p3] "<claim>" # SURVIVED / KILLED / UNRESOLVED
node <xllm-review.js> council p1,p2[,p3] "<question>" [--tiebreak] [--ready=a,b,c]
node <xllm-review.js> stats # pairwise agreement (measured decorrelation)
node <xllm-review.js> outcome <run-id> --adopted <spec|majority|minority|none> --helpful yes|no
| mode | epistemology | cost | reach for it when |
|---|---|---|---|
| roles | coverage — synthesis labels, not measured | 1× | advisors need different prompts |
| blind | measurement — identical blind prompt, append-only ledger | ~1× | you want the measured spread |
| debate | adversarial — decisive falsifiers kill wrong claims | ~2–3× | being wrong is expensive |
| council | independent divergence → hostile convergence | ~3–4× | the highest-stakes calls |
Reviewing code
Add exactly one diff source to any mode. The diff is collected deterministically (plain git), size-capped, sent to advisors — and never persisted (ledger/index record source/stat/bytes only):
--staged | --base <ref> | --diff-file <path>
--base <ref> is two-dot (git diff <ref>): working tree vs ref, so
uncommitted changes are included — NOT a merge-base (three-dot) branch diff.
Contract
- Advisors are read-only; your own vendor is refused; nothing is auto-applied.
- blind/council write
<state>/panel-ledger.jsonlBEFORE prose. The ledger is truth; your summary is UX; minority reports are findings; failures abstain. - roles output is never "measured" — its labels are your synthesis, not agreement rates. Only blind/council/stats speak measured decorrelation.
- On a blind split the core suggests the measured tiebreaker for free — spend
it with
--tiebreak; never hand-pick by vendor pedigree. - Afterwards record adoption:
outcome <run-id> …— it feeds measured routing. - SURVIVED = withstood refutation, not proven. Re-verify consequential claims.
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 · 70 lines · 102 tokens per session scan A df90027bc622
review is a skill published in the GitHub repository kimmingul/xllm (2 stars, last pushed 15d ago), licensed MIT. It adds 102 tokens to every session and 838 once invoked, about $0.0005 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
improve
Autonomous quality improvement loop. Scores a target against a rubric, selects the highest-leverage axis, attacks it, verifies, documents, and loops. No pre-planning between iterations — each loop re-scores from scratch.
triage
GitHub issue and PR investigator. Pulls open issues/PRs, classifies them, searches the codebase for root cause or reviews contributed code, proposes fixes with file:line references, and optionally implements fixes. Use for investigating GitHub issues and reviewing PRs; do NOT use for general code review unrelated to…
review
5-pass structured code review — correctness, security, performance, readability, consistency.
arch-review
Team baseline for architecture, design-doc, and RFC review — produces summary, strengths, concerns, and open questions. Auto-triggers when you say review this design, review this architecture, review this proposal, review my RFC, poke holes in this proposal, is this design sound, or critique this tech spec. Defers to…
security-review
Team baseline for security review — a best-effort heuristic security pass over a diff or change-set. Auto-triggers for review this for security, any vulnerabilities here, security pass on this diff, is this code secure, check this for security issues, or security review this. Uses installed scanners when available…
pr-review
Team baseline for PR and code review — produces scope, blockers, nits, and strengths. Auto-triggers when you say review this PR, check this diff, review these changes, pre-review before I share, what do you think of this PR, or review the diff on this branch. Defers to any other installed skill whose description…