aigne-afs-evaluation

aigne-afs-evaluation is an agent for Claude Code from EndogenAI/dogma. It costs 0 tokens per session (3,069 once invoked), scanned A, original, Apache-2.0.

An evaluation of AIGNE AFS, a local context-management system that stores agent information in a database and vector store and connects through MCP.

In plain words
What is it for?
Use it to assess context governance and local memory infrastructure for AI-agent fleets.
Why use it?
It examines whether a structured local store could reduce reliance on scratchpad files and manage which context agents keep, retrieve, compress, or isolate.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: mentions AGENTS.md.

Not installable on its own: it runs a file from its repository that does not travel with it. Clone the repository, or install whatever ships that file. The line is uv run python scripts/retrieve_afs_context.py \.

Install

Getting it into your agent

There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.

Made for: Claude Code.

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 aigne-afs-evaluation

README.md
[![agentmods](https://agentmods.dev/badge/agents/endogenai/dogma/aigne-afs-evaluation.svg)](https://agentmods.dev/agents/endogenai/dogma/aigne-afs-evaluation)
Your own site
<a href="https://agentmods.dev/agents/endogenai/dogma/aigne-afs-evaluation"><img src="https://agentmods.dev/badge/agents/endogenai/dogma/aigne-afs-evaluation.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,069 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00000 $0.03069
Opus 5 $0.00000 $0.01535
Sonnet 5 $0.00000 $0.00614
Haiku 4.5 $0.00000 $0.00307

Measured 2d ago against content hash df2ac4a69639, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

aigne-afs-evaluation 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 2d 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.

docs/research/agents/aigne-afs-evaluation.md · 220 lines

How it starts

The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AIGNE AFS Context Governance Layer Evaluation

Status: Final Research Question: Can the AIGNE Agent File System (AFS) context governance pipeline serve as the substrate for EndogenAI's token offloading strategy — replacing or augmenting the current scratchpad approach? Date: 2026-03-09


1. Executive Summary

The AIGNE AFS (Agentic File System) is a context governance layer backed by SQLite + a local vector store, exposing an MCP tool interface for agent integration. It implements a formal context engineering pipeline (Context Constructor → Context Updater → Context Evaluator) layered on top of the LangChain-attributed write→select→compress→isolate four-stage pattern (docs/research/agentic-research-flows.md §Attribution Correction).

Governing axioms: MANIFESTO.md §2 — Algorithms Before Tokens (a deterministic context manager reduces token burn per session) and MANIFESTO.md §3 — Local Compute-First (SQLite + local vector store, no cloud dependency).

Recommendation: MONITOR — do not adopt yet.

AIGNE AFS is architecturally the strongest candidate for a context governance layer among options surveyed. Its design directly addresses EndogenAI's primary context management gap: no semantic retrieval layer over episodic/experiential memory between sessions. However, it shares the same prerequisite as the episodic memory libraries evaluated in docs/research/episodic-memory-agents.md: the local compute baseline (OPEN_RESEARCH.md item 1) must be resolved before a local AFS deployment is viable. Premature adoption before that baseline is stable adds infrastructure complexity against an uncertain inference backend.

The current scratchpad-based approach remains the correct architecture for now. The evaluation below provides a concrete integration design that should be executed once the prerequisite is met.


2. Hypothesis Validation

H1 — The AIGNE AFS pipeline directly addresses the EndogenAI context management gap

Validated. The AIGNE paper enumerates seven memory types. The current EndogenAI architecture satisfies two of them:

Read the full file on GitHub · 220 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. 2d ago First seen · 220 lines · 0 tokens per session scan A df2ac4a69639

Subscribe to this mod's changes

aigne-afs-evaluation is an agent published in the GitHub repository EndogenAI/dogma (2 stars, last pushed 12d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,069 tokens. 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 agents, from other repositories

learning

The Learning agent is NXTG-Forge's institutional memory. It observes how you work -- which agents you invoke, what corrections you make, what coding patterns you prefer -- and encodes those observations into persistent preferences that improve every future session.

nxtg-ai/forge-plugin · 0 tokens

history-management

The runtime keeps conversation history for each agent session and sends a provider-facing working history to the model. Two complementary limits operate on different representations.

zeroclaw-labs/zeroclaw · 0 tokens

design-reviewer

Design-review sub-agent. Reviews design docs across the four dimensions of architecture, interface, performance, and security, covering MemOS's multi-memory / multi-storage backend constraints.

MemTensor/MemOS · 38 tokens

agent-system

An AgentSystem composes multiple Agents into a directed graph that Orloj executes as a coordinated workflow. The graph defines how messages flow between agents during task execution.

OrlojHQ/orloj · 0 tokens

compliance

The Compliance agent protects your project from legal and regulatory risk. It audits dependency licenses for compatibility (is that GPL-3.0 package going to force your MIT project open-source?), checks data handling against GDPR requirements (do you have a data inventory? can users request deletion?), verifies WCAG…

nxtg-ai/forge-plugin · 0 tokens

guardian

The Guardian is the last line of defense between your code and production. It runs a multi-dimensional quality gate that checks tests, type safety, security, code quality, and documentation coverage in a single pass. Unlike a CI pipeline that tells you something broke 10 minutes after you pushed, the Guardian catches…

nxtg-ai/forge-plugin · 0 tokens