context-firewall

context-firewall is a skill for Claude Code, Codex from jmagly/aiwg. It costs 31 tokens per session (567 once invoked), scanned A, original, MIT.

A read-only security review of the context that can enter AI provider sessions, including trust, changes, poisoning signals, and budget use. It can also prepare a reviewed baseline plan.

In plain words
What is it for?
Use it to scan context, check budget attribution, inspect a proposed memory baseline, or review entries before explicitly approving a baseline.
Why use it?
It helps identify suspicious, stale, changed, external, or over-budget context before it is used by an AI provider.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to scan context, check budget attribution, inspect a proposed memory…

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/jmagly/aiwg/context-firewall
Install

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.

Any agent
npx skills add jmagly/aiwg --skill context-firewall
Clone the repo
git clone --depth 1 https://github.com/jmagly/aiwg

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for context-firewall

README.md
[![agentmods](https://agentmods.dev/badge/skills/jmagly/aiwg/context-firewall.svg)](https://agentmods.dev/skills/jmagly/aiwg/context-firewall)
Your own site
<a href="https://agentmods.dev/skills/jmagly/aiwg/context-firewall"><img src="https://agentmods.dev/badge/skills/jmagly/aiwg/context-firewall.svg" alt="Measured on agentmods" height="20"></a>
Per session 31 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 567 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00031 $0.00567
Opus 5 $0.00015 $0.00283
Sonnet 5 $0.00006 $0.00113
Haiku 4.5 $0.00003 $0.00057

Measured 3d ago against content hash 0e43ef4106b6, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

context-firewall 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 3d 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.

agentic/code/addons/aiwg-utils/skills/context-firewall/SKILL.md · 60 lines

How it starts

The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Context Firewall

Use the public AIWG CLI to inspect context that can enter provider sessions. Keep scans read-only, never print source bodies as part of a threat finding, and treat baseline confirmation as an attestation that every planned record was reviewed.

Triggers

  • "audit context" → run the read-only firewall scan
  • "check context budget" → scan and explain budget attribution
  • "review memory baseline" → produce the full baseline plan without writing
  • "approve the context baseline" → plan first, then request explicit write authorization
  • "doctor says the context baseline is missing" → route to the plan-first workflow

Process

  1. Run aiwg context-firewall scan, adding repeatable --provider <name>, --strict, or --json only when the task needs them.
  2. Report violations and warnings using paths, trust labels, review states, digests, sizes, and signal identifiers. Do not reproduce suspicious content.
  3. If a baseline is requested, run aiwg context-firewall baseline --plan. Inspect every emitted record and surface any quarantined, stale, external, or changed entry before considering a write.
  4. Do not infer approval from a request to scan, diagnose, plan, or repair. Baseline mutation requires explicit user authorization after the plan has been presented and reviewed.
  5. Only after that authorization, run aiwg context-firewall baseline --write --confirm-reviewed. Use --output <project-relative-path> only when the user requested a custom destination.
  6. Re-run aiwg context-firewall scan --strict and report the resulting status.

The writer fails closed for unsafe records, paths outside the project root, symlink escapes, and missing confirmation. Do not bypass those checks or invoke the underlying implementation script directly.

Output

Report:

  • providers and total portable context budget;
  • attribution by memory, rule, skill, agent, generated bridge, and project-local;
  • trust and review findings by path, without source bodies;
  • whether the operation was read-only or changed the reviewed baseline;
  • the final strict scan status.

Read the full file on GitHub · 60 lines

Changes

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.

  1. 3d ago First seen · 60 lines · 31 tokens per session scan A 0e43ef4106b6

Subscribe to this mod's changes

context-firewall is a skill published in the GitHub repository jmagly/aiwg (209 stars, last pushed yesterday), licensed MIT. It adds 31 tokens to every session and 567 once invoked, about $0.0002 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 2026-09-03.

Related

Other skills, from other repositories

workflow-state-manager

Manages workflow state persistence and session recovery for sp.autonomous. Tracks phase progress, feature completion, and provides resume capability. Triggers: workflow state, session recovery, resume autonomous, progress tracking.

Wania-Kazmi/claude-code-autonomous-agent-workflow · 43 tokens

claudeception

Claudeception is a continuous learning system that extracts reusable knowledge from work sessions. Triggers: (1) /claudeception command to review session learnings, (2) "save this as a skill" or "extract a skill from this", (3) "what did we learn?", (4) After any task involving non-obvious debugging, workarounds, or…

Wania-Kazmi/claude-code-autonomous-agent-workflow · 99 tokens

unified-memory

Share durable, inspectable context and handoffs between Claude, Codex, Hermes, Cursor, OpenCode, and other agents through the local ECC Memory Vault. Use when an agent must save work state, transfer context, resume another agent's task, or search shared project knowledge.

affaan-m/ECC · 60 tokens

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…

cwinvestments/memstack · 79 tokens

state

Use when the user says 'update state', 'project state', 'where was I', or at session start to load current context.

cwinvestments/memstack · 29 tokens

Vizra ADK Memory System

Implement persistent memory, session context, and vector memory (RAG) for AI agents.

vizra-ai/vizra-adk · 24 tokens