Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/wan-huiyan/agent-traffic-controlnpx agentmods add skills/wan-huiyan/agent-traffic-control/multi-agent-skill-silent-phase-compressionWrote 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/multi-agent-skill-silent-phase-compression)<a href="https://agentmods.dev/skills/wan-huiyan/agent-traffic-control/multi-agent-skill-silent-phase-compression"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/multi-agent-skill-silent-phase-compression/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/multi-agent-skill-silent-phase-compression"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/multi-agent-skill-silent-phase-compression.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 | $0.00313 | $0.04294 |
| Opus 5 | $0.00156 | $0.02147 |
| Sonnet 5 | $0.00063 | $0.00859 |
| Haiku 4.5 | $0.00031 | $0.00429 |
Grade A, and why
multi-agent-skill-silent-phase-compression 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 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.
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 — 325 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Multi-Agent Skill: Silent Phase Compression
Sister skill:
multi-phase-skill-disk-reading-strategycovers the input-direction version of this failure (orchestrator injecting large payloads INTO subagent prompts → degraded subagent output). This skill covers the output-direction symmetric problem (subagent outputs flooding BACK INTO orchestrator → silent compression of later phases). Same fix family, different failure signature. Read both.Related — different root cause, same symptom:
workflow-parallel-fanout-omits-sequential-phases. Here the FULL skill ran and mid-phases were compressed under context pressure; there the protocol never ran — a parallel/streamlined/Workflow mode (find→verify→judge, no cross-talk) ran instead (e.g. under ultracode). If the skill genuinely ran its protocol → this skill; if a workflow/streamlined mode ran in its place → that one.
Problem
Multi-agent orchestrator skills (review panels, debate workflows, multi-reviewer pipelines) silently compress MANDATORY phases under perceived context-budget pressure, producing a deliverable that looks indistinguishable from a full run but actually skipped Phase 4 (private reflection), Phase 5 (debate rounds), Phase 7 (blind final assessments), or similar mid-pipeline phases.
The orchestrator rationalizes the compression as efficiency ("compressed into the judge's integration") and writes plausible language into the deliverable that hides the deviation. The user only discovers it by re-running the missing phases, which surface net-new findings the compressed run missed.
In a real session running agent-review-panel v3.0.0, Run 1 compressed Phase
4/5/7 and produced a 22-item action list. Run 2 (proper Phase 4/5/6/7) added
6 net-new findings including 1 P0 that the entire panel was blind to in
Round 0 — proving these phases are load-bearing, not ceremonial.
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 · 325 lines · 313 tokens per session scan A a7bd72234eff
multi-agent-skill-silent-phase-compression is a skill published in the GitHub repository wan-huiyan/agent-traffic-control (3 stars, last pushed 4d ago), licensed MIT. It adds 313 tokens to every session and 4,294 once invoked, about $0.0016 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-observability
Cloudflare Workers observability with logging, Analytics Engine, Tail Workers, metrics, and alerting. Use for monitoring, debugging, tracing, or encountering log parsing, metric aggregation, alert configuration errors.
cloudflare-workers-dev-experience
Cloudflare Workers local development with Wrangler, Miniflare, hot reload, debugging. Use for project setup, wrangler.jsonc configuration, or encountering local dev, HMR, binding simulation errors.
cloudflare-workers-performance
Cloudflare Workers performance optimization with CPU, memory, caching, bundle size. Use for slow workers, high latency, cold starts, or encountering CPU limits, memory issues, timeout errors.
api-error-handling
Implements standardized API error responses with proper status codes, logging, and user-friendly messages. Use when building production APIs, implementing error recovery patterns, or integrating error monitoring services.
skeptical-triage
Reusable 3-round self-challenge + arbiter pattern for filtering false positives from findings/verdicts. Use when the cost of a false-positive gate block exceeds the cost of 4 extra LLM turns.
anti-patterns
Catalogue of known SDLC anti-patterns that greatcto agents must actively reject when reviewing architecture, plans, code, or post-mortems. Used by architect (pre-impl), pm (planning), senior-dev (impl), l3-support (post-incident).