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 alinaqi/maggy --skill context-shuntgit clone --depth 1 https://github.com/alinaqi/maggyWrote 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/alinaqi/maggy/context-shunt)<a href="https://agentmods.dev/skills/alinaqi/maggy/context-shunt"><img src="https://agentmods.dev/badge/skills/alinaqi/maggy/context-shunt/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/alinaqi/maggy/context-shunt"><img src="https://agentmods.dev/badge/skills/alinaqi/maggy/context-shunt.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 | $0.00028 | $0.00574 |
| Opus 5 | $0.00014 | $0.00287 |
| Sonnet 5 | $0.00006 | $0.00115 |
| Haiku 4.5 | $0.00003 | $0.00057 |
Grade A, and why
context-shunt 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.
What it actually says
Context Shunt — Read Cheap, Keep Context Small
Reading big files into context is the most expensive thing an agent does for the
least reasoning value. The shunt hands those reads to a cheap worker model
(bulk-read) that answers a question about the files and returns a compact
summary. The raw bytes never enter this session's context.
This is orthogonal to whole-turn routing (srooter / route-task): those pick the
model for the turn; the shunt trims what a tool call pulls into context when
the turn is legitimately here.
Decision: read raw, shunt, or graph?
- Editing this exact file — read it raw. You need every line; never edit against a summary.
- A fact/answer across large or many files —
bulk-read "<question>" file.... - A code symbol (function/class/route) —
get_code_snippet(qualified_name): free and exact. - Small file (under threshold) you need in full — read it raw.
A shunt answer is for understanding, not for producing a diff.
Usage
bulk-read "how does token refresh work?" src/auth/session.ts src/auth/refresh.ts
bulk-read "which config keys are read at startup?" $(git ls-files 'config/*.yaml')
bulk-read prints structured bullets citing path:line, or
NOT FOUND IN PROVIDED FILES. A token-savings report goes to stderr.
Configuration
Env vars or ~/.claude/shunt.conf (see templates/shunt.conf):
SHUNT—on/offmaster switch for the PreToolUse hook.SHUNT_MIN_LINES— large-read threshold (default 350).SHUNT_MODE—suggest(default) /block/offfor the hook's large-read action.SHUNT_GRAPH_NUDGE—on/offonce-per-session graph nudge.SHUNT_MODEL— worker command (defaultdeepseek --flash; alsogemini-api --flash-lite,qwen3,glm).
The context-shunt-gate PreToolUse hook enforces the thresholds; this skill tells
you when to reach for bulk-read yourself.
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 First seen · 51 lines · 28 tokens per session scan A e6ed6f09b6ec
context-shunt is a skill published in the GitHub repository alinaqi/maggy (706 stars, last pushed 2d ago), licensed MIT. It adds 28 tokens to every session and 574 once invoked, about $0.0001 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 2026-09-09.
Other skills, from other repositories
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.
memstack-automation-n8n-workflow-builder
Use this skill when the user says 'n8n workflow', 'build a workflow', 'automation workflow', 'connect services', or needs visual workflow design with node mapping, data transformations, and error handling for n8n. Do NOT use for standalone webhook endpoints or cron jobs.
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…
state
Use when the user says 'update state', 'project state', 'where was I', or at session start to load current context.
grimoire
Use when the user says 'update context', 'update claude', 'save library', or after significant project changes.
save-progress
Save current project task progress to local task state file for cross-session continuity. Use when the user asks to save progress, uses an equivalent localized trigger phrase, the session is interrupted, or the user wants to resume work later. Writes to .claude/project-task-state.json so next session can load it via…