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 wan-huiyan/agent-traffic-control --skill pre-dispatch-agent-isolation-parameter-not-promptgit clone --depth 1 https://github.com/wan-huiyan/agent-traffic-controlWrote 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/wan-huiyan/agent-traffic-control/pre-dispatch-agent-isolation-parameter-not-prompt)<a href="https://agentmods.dev/skills/wan-huiyan/agent-traffic-control/pre-dispatch-agent-isolation-parameter-not-prompt"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/pre-dispatch-agent-isolation-parameter-not-prompt/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/wan-huiyan/agent-traffic-control/pre-dispatch-agent-isolation-parameter-not-prompt"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/pre-dispatch-agent-isolation-parameter-not-prompt.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 | $0.00075 | $0.02477 |
| Opus 5 | $0.00037 | $0.01239 |
| Sonnet 5 | $0.00015 | $0.00495 |
| Haiku 4.5 | $0.00007 | $0.00248 |
Grade A, and why
pre-dispatch-agent-isolation-parameter-not-prompt 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 5d 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.
Agent isolation is a parameter, not a sentence in the prompt
Problem
You are about to fan out several agents that will each commit to their own branch. Your prompt template opens with "You are running in an isolated git worktree — changes you make here do NOT affect other agents", so you dispatch and move on.
That sentence creates nothing. The worktree is created by the dispatch tool's
isolation: "worktree" parameter. With the parameter unset, every agent runs in the same
checkout — and every one of them reports back as though it were isolated, because you told it
it was.
Nothing errors. Git does not complain that four agents share a working tree; that is a perfectly ordinary thing for a directory to be. The damage arrives as ordinary-looking results: a commit on the wrong branch, a branch that changes under a session mid-run, a test total that is off by one.
Why this specific mistake keeps happening: the isolation is a field in a tool call, and a field is very easy to describe in English and forget to set. The prompt is the part you write by hand and re-read; the parameters are the part you fill in once from a template. So the belief that the agents are isolated is strongest precisely when it is false.
Context / Trigger Conditions
Any one of these is enough — this is a cheap check, not a gated procedure:
- You are about to issue 2 or more Agent / Workflow / Task calls in one batch, and the
agents will run
git commit,git checkout,git rebaseorgit push. - Your dispatch prompt asserts isolation ("your own worktree", "changes here don't affect others") — that phrasing is the tell that you are relying on prose.
- You are reusing a prompt template or a plan document that was written for an earlier fan-out, where the parameters are not part of the text you copied.
- A single agent is being sent into a directory another session is live in. One agent is enough; "parallel" is not a precondition.
Not for: read-only agents that never write to the repo (a review pass, a grep audit).
Those have a different hazard — reading the wrong tree — which
subagent-read-stale-worktree-needs-head-pin covers.
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.
- 5d ago Changed · +7 lines b832a632677e
- 10d ago First seen · 170 lines · 75 tokens per session scan A f7936a9d5ba7
pre-dispatch-agent-isolation-parameter-not-prompt is a skill published in the GitHub repository wan-huiyan/agent-traffic-control (3 stars, last pushed 5d ago), licensed MIT. It adds 75 tokens to every session and 2,477 once invoked, about $0.0004 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
cloudflare-workers-ci-cd
Complete CI/CD guide for Cloudflare Workers using GitHub Actions and GitLab CI. Use for automated testing, deployment pipelines, preview environments, secrets management, or encountering deployment failures, workflow errors, environment configuration issues.
claude-code-bash-patterns
Claude Code Bash tool patterns with hooks, automation, git workflows. Use for PreToolUse hooks, command chaining, CLI orchestration, custom commands, or encountering bash permissions, command failures, security guards, hook configurations.
bump-sdk-version
Bump the FutureSearch SDK version across all files. Use when releasing a new SDK version, updating version numbers, or the user says bump version, release, version bump.
pr-check
PR review compliance: fetch review comments from an open PR, categorize as resolved/unresolved/dismissed, critically evaluate fixable items, implement approved fixes, and reply inline. Pass --unattended to halt on human-judgment items (emitted as Pending-Human) instead of prompting via AskUserQuestion.
babysit
PR babysitter: monitors CI status, auto-rebases when behind, auto-fixes CI where possible, delegates review comment handling to dlc:pr-check, and re-requests review after fixes. Designed for /loop usage with Remote Control.
git-ops
Git hygiene: clean up merged branches, prune stale remotes, and verify repository state.