Ross ROSS = Recommend OSS · open-source software intelligence for agents

FailproofAI/failproofai

Observability and enforcement for AI agent harnesses. Capture every run and runtime reliability with policy enforcement. 40 built-in policies, a local dashboard, no account required with a generous free cloud plan observed · 2026-08-28

github.com/FailproofAI/failproofai · homepage · MDX · NOASSERTION (other) observed · 2026-08-28

Health v2 · maintenance only

81/100

  • Activity 99
  • Release rhythm 98
  • Longevity 10

Flags: young no_license

How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.

  • gap_med: 0.0
  • age_days: 150
  • days_rel: 14
  • days_push: 7
  • n_releases_24m: 27

Full methodology

Adoption not part of the score

1508 stars · 404 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Failproof AI is an open-source observability and enforcement layer for AI agent harnesses like Claude Code, Codex, Hermes, and OpenClaw, capturing every run and blocking dangerous tool calls in real time via hooks and a sub-millisecond Rust daemon. It ships as an npm CLI with a local dashboard, 40 built-in policies, a Python SDK for custom agents, and an optional cloud plan for teams.

Use cases

  • observe and trace every AI agent run, tool call, and model call
  • block dangerous tool calls like rm -rf or force pushes before they execute
  • detect agent failure modes like infinite loops, hallucinated tools, and context drift
  • audit agent sessions and cluster recurring failures into ranked issues
  • add guardrails and policy enforcement to coding CLIs and chat gateways
  • run agent reliability monitoring locally with no account
  • add observability to unattended agents running behind chat gateways or cron

When to choose

  • you run agents in supported harnesses (Claude Code, Codex, Cursor, Gemini CLI, Hermes, OpenClaw, etc.) and need runtime enforcement, not just logging
  • you want local-first tracing and auditing with a free tier and no account required
  • you need fail-closed policy enforcement with sub-millisecond latency in the tool-call path

When to avoid

  • your agent runs in an unsupported runtime and you need enforcement, not just SDK-based tracing
  • you need a fully permissive MIT license for commercial redistribution (it is MIT + Commons Clause)
  • you need Windows support today (setup supports Linux and macOS only)

Facets

cli-tool · maturity active

monitoring tracing alerting logging security cli agent-framework analytics large-language-models developer-tools monitoring security cli self-hosted cloud agent-observability policy-enforcement guardrails claude-code codex agent-harness llm-observability evals rust-daemon fail-closed ai-agents automation linux macos nodejs docker

6 sources

Member repositories

RepositoryRoleHealth v2
FailproofAI/failproofaimain81

For agents

markdown · JSON · MCP: product_card(name="FailproofAI/failproofai")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem