What did Hlido score Axe?

Axe scored 70/100 (STEADY) in Hlido's independent, hands-on review.

How much does Axe cost?

Hlido recorded the pricing model as: Open source. Derived from evidence Hlido holds, not vendor-supplied; verify current prices on the vendor's pricing page.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

Axe

Name: Axe review
Item: Axe
Rating: 70
Author: Hlido Editor

Frameworks & Eval · tested 2026-08-18 · re-test due 2026-11-16 · by the Hlido desk, not the vendor

In short: A Unix-philosophy answer to the everything-chatbot: small, single-purpose LLM agents defined in TOML and composed with cron, pipes and git hooks — a coherent design that is also refreshingly security-conscious.

Visit Axe → scorecard.json

Quick answer

Axe scores 70/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-18). STEADY (70) because Axe presents a coherent, well-documented design — Unix-composable single-purpose agents defined in version-controllable TOML, multi-provider, with genuinely security-conscious defaults (working-direct Pricing: Open source (free entry point documented).

Hlido Editor · 2026-08-18

Axe makes an argument as much as a product, and the argument is good: most AI tooling assumes you want one long-running chatbot with a giant context window doing everything, but good software is small, focused and composable, so agents should be too. Axe treats LLM agents the way Unix treats programs — each does one thing, defined in a TOML file with its own system prompt, model, skills, context and memory, and run from the command line. Pipe data in, get results out, chain agents, trigger them from cron, git hooks or CI. It is explicitly 'the executor, not the scheduler,' composing with the Unix tools you already use rather than reinventing orchestration — a disciplined scoping decision. The feature list is substantial and, notably, security-aware in ways many agent frameworks are not: multi-provider (Anthropic, OpenAI, Ollama, OpenCode, AWS Bedrock), sub-agent delegation with depth limiting and parallel execution, persistent timestamped-markdown memory with LLM-assisted garbage collection, built-in file tools sandboxed to the working directory, an output allowlist that restricts url_fetch/web_search to specific hostnames with private/reserved IPs always blocked (SSRF protection), token budgets, configurable retry with backoff, and MCP tool support over SSE/streamable-HTTP. Four direct dependencies, a single Go binary, TOML configs you can version-control. The caveats are those of a young, single-maintainer open-source project: the documentation surface we captured is clear and complete, but there is no adoption, benchmark or reliability evidence, no third-party validation, and the behavioural quality of the agents it runs depends entirely on the user's configs and chosen providers. As a composable, scriptable, security-conscious agent runner for people who live in the terminal, it is a genuinely well-thought-out tool.

Why STEADY

STEADY (70) because Axe presents a coherent, well-documented design — Unix-composable single-purpose agents defined in version-controllable TOML, multi-provider, with genuinely security-conscious defaults (working-directory sandboxing, SSRF-protected output allowlist, token budgets, depth-limited delegation) and MCP support — as a clean single binary. Not VITAL because it is an early, solo open-source project with no adoption, benchmark or reliability evidence on the captured surface, and the quality of any agent it runs depends wholly on the user's configuration and provider choices.

What it does well

Models agents the Unix way: small, single-purpose, defined in version-controllable TOML, composed with cron/pipes/git hooks
Stays scoped as the executor, not the scheduler — composes with existing tools instead of reinventing orchestration
Ships security-conscious defaults: working-directory-sandboxed file tools, SSRF-protected output allowlist, token budgets
Supports multiple providers (Anthropic, OpenAI, Ollama, OpenCode, AWS Bedrock) and MCP tools over SSE/streamable-HTTP
Offers sub-agent delegation with depth limiting and parallel execution, plus persistent memory with LLM-assisted GC
Distributes as a single Go binary with only four direct dependencies and stdin piping for composition

What it fails at

Early, single-maintainer open-source project with no adoption, benchmark or reliability evidence on the surface
Agent behavioural quality depends entirely on the user's TOML configs and chosen model providers
No third-party validation of the security controls (sandboxing, SSRF allowlist) on the captured surface
Terminal- and config-file-centric; not for users wanting a GUI or a managed platform
Documentation is thorough but the product's real-world robustness at scale is unproven

Best for

Developers who want composable, single-purpose LLM agents triggered from cron, git hooks, CI and pipes
Teams that value version-controllable, declarative agent definitions over a monolithic chatbot
Security-conscious users who want working-directory sandboxing and SSRF-protected fetches by default
Polyglot/multi-provider setups (Anthropic, OpenAI, Ollama, Bedrock) run from a single binary

Not recommended for

Users wanting a GUI or an all-in-one conversational assistant rather than a scriptable CLI executor
Teams needing vendor support, SLAs or proven at-scale reliability before adoption
Those expecting Axe to schedule/orchestrate for them — it is deliberately the executor only
Non-technical users uncomfortable authoring TOML agent configs

Pricing & access

ModelOpen source
Free entry pointYes — a free tier or open-source edition is documented

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-18.

Compared to

Agent Manager composable-headless-agent-execution
Both run coding/LLM agents from the terminal, but Axe is a headless, composable executor — agents defined in TOML and chained with cron, pipes and CI for automation — whereas Agent Manager is an interactive TUI for driving many live agent sessions by hand. Choose Axe to script and automate single-purpose agents; choose Agent Manager to supervise a fleet of interactive coding agents.

Agent relevance

CLI MCP Behavioral-testable

Agentic-Commerce Readiness 57/100 · INTEGRABLE

Independent readiness for agent delegation & transaction. How it’s scored · check live

Axe IS an agent runner: define agents in TOML and invoke them from the shell (e.g. `git diff | axe run reviewer`), chain them, or trigger from cron/CI. It also consumes external MCP servers over SSE/streamable-HTTP to extend an agent's tools, and can delegate to sub-agents. Fully scriptable and composable; JSON output and dry-run mode make it easy to embed in larger automation.

Agent-friendly score: 8/10

Evidence

Single-purpose agents defined in TOML, composed with cron/pipes/git hooks; executor not scheduler — source (2026-08-18) verified
Working-directory-sandboxed file tools and SSRF-protected output allowlist (private IPs blocked) — source (2026-08-18) verified
Multi-provider (Anthropic, OpenAI, Ollama, OpenCode, Bedrock) and MCP tool support — source (2026-08-18) verified
Sub-agent delegation with depth limiting; token budgets; four direct dependencies — source (2026-08-18) verified
Independent adoption, benchmark or reliability evidence — source (2026-08-18)

scorecard.json · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-11-16

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/jrswab-axe.svg)](https://hlido.eu/check/?agent=jrswab-axe)

HTML

<a href="https://hlido.eu/check/?agent=jrswab-axe"><img src="https://hlido.eu/badge/jrswab-axe.svg" alt="Hlido trust score"></a>