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/Pinperepette/context-kernelWrote 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/plugins/pinperepette/context-kernel/claude-context-kernel)<a href="https://agentmods.dev/plugins/pinperepette/context-kernel/claude-context-kernel"><img src="https://agentmods.dev/badge/plugins/pinperepette/context-kernel/claude-context-kernel/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/plugins/pinperepette/context-kernel/claude-context-kernel"><img src="https://agentmods.dev/badge/plugins/pinperepette/context-kernel/claude-context-kernel.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
context-kernel 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 1mo 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 — 18 lines — stays where its author put it; the contents beside it link to each section on GitHub.
{
"name": "context-kernel",
"description": "Pipeline P_Q di proiezione del contesto: T1 compressione signal-preserving degli output dei tool (canary end-to-end con auto-degrade: dopo N violazioni updatedToolOutput nella sessione, raw pass-through; auto-heal su evidenza dal transcript: auto-ack dei failure transitori dopo M verified consecutive e un-degrade a sonda — 1 output comprimibile ogni K ripassa dal flusso normale, M sonde verificate consecutive tolgono il degrado, mai a tempo, code-aware sui sorgenti, JSON-aware sugli output MCP, page fault su elisione con recall mirato e recall storage di sessione (recall --search: grep su tutti gli output parcheggiati), escape per-comando # ck:raw), T2 slice deterministico del repository dal sintomo (budget automatico, AMBIENTALE: traceback nel prompt o test falliti nell'output Bash -> manifest iniettato; cambio-task dichiarato col diff dei manifest; resolver supervisionato dei riferimenti dinamici: import_module/__import__ con argomento letterale -> seed col call site, il resto punto cieco dichiarato, mai indovinato; slice formale a simbolo su budget insoddisfacibile: Python esatto via AST, Go conservativo via tokenizer con fallback al file intero su split infido; oracle di sufficienza deterministico: la proiezione e' sufficiente sse contiene la chiusura answer-preserving dei seed, altrimenti i file droppati dal budget sono dichiarati page fault attesi; grafo generico promosso a preciso dall'indice ctags se il repo lo shippa: archi simbolo->definitore univoco, zero dipendenze, ambiguo saltato; prior da accoppiamento evolutivo git: i file che co-cambiano coi seed nella storia diventano seed additivi, segnale cold-start ortogonale a T5; prior di copertura: consuma .coverage/coverage.xml/lcov.info se presenti (stdlib, zero dip) e aggiunge come seed i file ESEGUITI fuori dalla chiusura statica = reachability dinamica che il grafo perde, troppi -> dichiarati non seminati), T3 carta del task con vincoli citati — ATTIVA: guaWhat it installs
The manifest is a name and a version. 6 skills, 8 commands, 3 agents, 6 hooks, 1 MCP server travel with it, and installing the plugin installs all of them — 1,067 tokens a session between them. Each is measured on its own page, and each can be installed alone.
- Skill kernel-repo-slice A 84 tokens
- Skill kernel-invariants A 72 tokens
- Skill kernel-pipeline A 69 tokens
- Skill kernel-slice A 68 tokens
- Skill kernel-ops A 207 tokens
- Skill kernel-verify A 62 tokens
- Command ck-charter A 33 tokens
- Command ck-recall A 35 tokens
- Command ck-tune A 39 tokens
- Command ck-savings A 37 tokens
- Command ck-status A 33 tokens
- Command ck-verify A 36 tokens
- Command ck-doctor A 32 tokens
- Command ck-smoke A 28 tokens
- Agent kernel-extractor A 72 tokens
- Agent kernel-verifier A 76 tokens
- Agent kernel-scout A 84 tokens
- Hook PreCompact A not measured
- Hook SessionEnd A not measured
- Hook UserPromptSubmit A not measured
- Hook SessionStart A not measured
- Hook PreToolUse A not measured
- Hook PostToolUse A not measured
- MCP server context-kernel A not measured
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.
- 1mo ago First seen · 18 lines scan A d602c7b498f3
context-kernel is a plugin published in the GitHub repository Pinperepette/context-kernel (24 stars, last pushed 2mo ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 plugins, from other repositories
open-knowledge-vault
Create and maintain portable Open Knowledge Format vaults with optional Obsidian and Dataview views.
knowledge-and-memory
Research methodology, corpus search, persistent memory, iterative workflows, and project knowledge management.
pai
Personal AI Infrastructure — persistent memory, session continuity, and knowledge graph for Claude Code.
tree-ring-memory
Local-first memory lifecycle, project bootstrap, and receipt-backed harness guidance for Claude Code using Tree Ring Memory v0.15+.
claude-mega-brain
OKF-powered knowledge context for Claude Code — injects your project knowledge base at every session.
vault-template
Personal Knowledge Management system for Obsidian with AI-powered workflows, goal tracking, and productivity coaching.