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 TheColliery/CoalWash --skill coalwashgit clone --depth 1 https://github.com/TheColliery/CoalWashWrote 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/thecolliery/coalwash/coalwash)<a href="https://agentmods.dev/skills/thecolliery/coalwash/coalwash"><img src="https://agentmods.dev/badge/skills/thecolliery/coalwash/coalwash/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/thecolliery/coalwash/coalwash"><img src="https://agentmods.dev/badge/skills/thecolliery/coalwash/coalwash.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 | $0.00254 | $0.07411 |
| Opus 5 | $0.00127 | $0.03705 |
| Sonnet 5 | $0.00051 | $0.01482 |
| Haiku 4.5 | $0.00025 | $0.00741 |
Grade A, and why
coalwash 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 today.
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 — 191 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CoalWash — the memory washer
Fidelity scope — CLASS-B wash tiers only (class-A estate = the byte-identity contract two notes down): the gate PROVES every STRUCTURED token that went in came out (the classes at step 3) by mechanical diff; it CANNOT SEE survivors trading places (
pass 878/fail 0→pass 0/fail 878passes it) — re-pairing + load-bearing prose facts are the semantic reviewers' + YOUR job, never the gate's. Deletes ride the adjudicated plan, not a separate approval; safety is the transactional apply (snapshot → whole-run rollback; a rollback whose own restore fails reports partial, never a silent mixed state). CoalWash slows memory-overhead growth — it does not eliminate it.
[!CAUTION] Fat grows at a different rate per project — a schedule cleans lean memory, and past fat-exhaustion a semantic pass can throw away something load-bearing (prohibitions #1–2). A looped Full tier re-pays semantic cost every round for nothing, and accumulates over-compression pressure.
Class-A estate (ULTRA + RE-TIER's byte lane) — a DIFFERENT, STRONGER contract: transcripts/tool-results are vendor-owned + machine-parsed (fail the 4 wash tests) → NEVER semantic-edited; the guarantee is byte-identity (copy-verify-then-delete,
estate-restoreround-trips) — not the structured-token gate above. The seams where the class-B gate re-enters = RE-TIER's hot-index rewrite + choice 4's ③ manual-tier work (fidelity gate + MOVE-VERIFY + #54 anchor bind exactly there). Detail →references/method.md§10/§11.
You are the insider/orchestrator — the heavy core is CODE (the engine modules beside this skill); you do ONLY the semantic judgment a script cannot. Resolve LIB = ../../scripts/lib from this file (plugin and file-copy layouts identical; confirm fidelity-gate.mjs is there). Deep detail — snippets, the outsider rubric, the garbage taxonomy — is references/method.md; Claude Code adapter facts (paths, caps, state files) are references/platform-cc.md. Both load on-demand; the cheap path never pays for them.
What ships with it
2 files 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.
- today Changed · +1 tokens per session 80b40827fade
- 10d ago First seen · 191 lines · 253 tokens per session scan A eaf4ed9a3808
coalwash is a skill published in the GitHub repository TheColliery/CoalWash (11 stars, last pushed today), licensed Apache-2.0. It adds 254 tokens to every session and 7,411 once invoked, about $0.0013 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 2026-08-30.
Other skills, from other repositories
session-management
Context preservation, tiered summarization, resumability.
code-graph
AST-based code graph for fast symbol lookup, dependency analysis, and blast radius via codebase-memory-mcp MCP server.
mnemos
Task-scoped memory lifecycle — typed MnemoGraph prevents lossy context compaction by treating facts/decisions/code-refs/handoffs as distinct node types with per-type eviction policies.
context-shunt
Offload large / multi-file reads to a cheap worker model so raw files never enter Claude's context (token savings).
report
Writes the session final report to a file, then prints only the path and a one-line summary. Fires when the prompt contains "Report per memstack:report", and also when the prompt begins with a standing trigger configured through MEMSTACKREPORTONTASKPROMPTS or MEMSTACKREPORTTRIGGERS. Dormant otherwise.
token-optimization
Use when the user says 'token optimization', 'save tokens', 'context window', 'reduce tokens', 'token stack', or 'TokenStack', or asks about extending context window capacity. Covers TokenStack, the built-in compression proxy that shrinks Claude Code tool output before it reaches the Anthropic API. Do NOT use for…