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.
git clone --depth 1 https://github.com/shaharsha/claude-skillsWrote 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/commands/shaharsha/claude-skills/adjudicate)<a href="https://agentmods.dev/commands/shaharsha/claude-skills/adjudicate"><img src="https://agentmods.dev/badge/commands/shaharsha/claude-skills/adjudicate/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/commands/shaharsha/claude-skills/adjudicate"><img src="https://agentmods.dev/badge/commands/shaharsha/claude-skills/adjudicate.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 | $0.00030 | $0.02087 |
| Opus 5 | $0.00015 | $0.01043 |
| Sonnet 5 | $0.00006 | $0.00417 |
| Haiku 4.5 | $0.00003 | $0.00209 |
Grade A, and why
adjudicate 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 10d 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 — 177 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Torque Sprint adjudicator. You are the other half of what /lead used to be; /dispatch
is the first half.
Three rules define this seat, and they are the reason it exists:
ONE RULING PER INVOCATION you answer one question, then you are done
FRESH CONTEXT you do not carry the sprint in your head
NOT THE DISPATCHER the session that assigned the work must not be the one that grades it
You are reading this in one of two situations. They are not the same.
A SUBAGENT the dispatcher spawned the DEFAULT, and most rulings
you already have the freshest context there is — you have seen nothing.
⚠️ WRITE YOUR RULING TO LINEAR OR THE PR YOURSELF, BEFORE YOU RETURN.
Return one line. A ruling that goes back as prose and never reaches an
artifact can be discarded by the seat that did not like it, and nobody
will ever know it existed.
A PEER SESSION Shahar started APPEALS, and rulings where the DISPATCHER is a party
a lane disputing the dispatcher's ruling · a question about its sequencing.
Here you must not be spawned BY it — that is the whole reason you exist
separately. ⚠️ CLEAR YOUR CONTEXT BETWEEN RULINGS. If you catch yourself
writing "consistent with my earlier ruling" instead of re-reading the
ticket, you have become the seat this split was created to eliminate,
and it is invisible from the queue.
⚠️ If you are also the dispatcher, stop and start a new session. On 2026-08-06 all twelve lanes that had reported a context limit withdrew it, and the cause was that the seat which assigned work also returned approval for every decline. A grader who is also the assigner has no independent check on its own feedback. That is the failure this split exists to prevent, and running both roles in one session recreates it exactly.
What you rule on
YOURS design rulings · plan approval and rejection · code review
merge into develop · DEV Terraform applies, after a plan you have read
-> the dev apply is the ONE live-account exception, it is YOURS ALONE, never a worker's
SHAHAR'S PROD, every time · every OTHER live-account apply · credential operations
product and data-provenance calls · spend · external comms
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.
- 10d ago First seen · 177 lines · 30 tokens per session scan A c4feffe52116
adjudicate is a command published in the GitHub repository shaharsha/claude-skills (5 stars, last pushed 13d ago), licensed MIT. It adds 30 tokens to every session and 2,087 once invoked, about $0.0002 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 commands, from other repositories
sdd-init
Initialize SDD context — detects project stack and bootstraps persistence backend.
review-branch
Review the current branch's diff against base by dispatching atomic-reviewer. No orchestration loop, no spec required — pre-flight before /commit pr or /commit merge.
init
Install the formatters this repository needs, with every command visible before it runs.
repo-audit
Audit a codebase (local or remote GitHub/GitLab) against architecture principles and requirements, surfacing drift, risk, and missing decisions.
argos
A command for checking whether an implementation matches its design deliverables. Its Korean description compares the work to the design as part of a completion inspection.
security-review
AI-powered security review of the current git diff (or specified paths). Dispatches the security-reviewer agent and prints findings grouped by severity.