agent47
llms.txt4 starsChanged 40 days agoMITView on GitHub
Python runtime guards for instrumented AI agents. Repository: agent47. Install agentguard47; import agentguard. The base SDK has no runtime dependencies. Checks apply only where integrated. Budget preflight checks recorded usage; it does not reserve concurrent capacity or predict the next response's cost. Streamed OpenAI and Anthropic calls record final usage once. Guards are not a sandbox. Local traces need no hosted account. Remote sinks are optional. The OpenAI Responses API, host tools, and provider invoices are unsupported. See enforcement-boundary.md.
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- next-ticketSkill
- agentguardSkill
The file
Files
# AgentGuard > Python runtime guards for instrumented AI agents. Repository: agent47. > Install agentguard47; import agentguard. The base SDK has no runtime dependencies. ## Read first - [README](https://github.com/bmdhodl/agent47/blob/main/README.md): Installation, runnable example, enforcement limits. - [Enforcement boundary](https://github.com/bmdhodl/agent47/blob/main/docs/enforcement-boundary.md): tested classes and remaining exposure. - [Documentation index](https://github.com/bmdhodl/agent47/blob/main/docs/README.md): Guides and references by task. - [Getting started](https://github.com/bmdhodl/agent47/blob/main/docs/guides/getting-started.md): Local trace and provider setup. - [Activation evidence](https://github.com/bmdhodl/agent47/blob/main/docs/guides/activation-metrics-design.md): Page views are not installs. Demo feedback is voluntary and local. - [Reservation contract](https://github.com/bmdhodl/agent47/blob/main/docs/guides/reservation-contract.md): Local reserve/commit/cancel/unresolved rules. Store-backed sync OpenAI Chat Completions only. Not an invoice cap. - [Contributing](https://github.com/bmdhodl/agent47/blob/main/CONTRIBUTING.md): Development setup and checks. - [Agent instructions](https://github.com/bmdhodl/agent47/blob/main/AGENTS.md): Repository contribution rules. ## Source references - [Package metadata](https://github.com/bmdhodl/agent47/blob/main/sdk/pyproject.toml): Branch version, Python floor, optional dependencies. - [Public API](https://github.com/bmdhodl/agent47/blob/main/sdk/agentguard/__init__.py): Supported imports. - [Guards](https://github.com/bmdhodl/agent47/blob/main/sdk/agentguard/guards.py): Budget, loop, timeout, retry, and rate checks. - [Changelog](https://github.com/bmdhodl/agent47/blob/main/CHANGELOG.md): Versioned changes. - [Published package](https://pypi.org/project/agentguard47/): Published metadata. ## Boundaries Checks apply only where integrated. Budget preflight checks recorded usage; it does not reserve concurrent capacity or predict the next response's cost. Streamed OpenAI and Anthropic calls record final usage once. Guards are not a sandbox. Local traces need no hosted account. Remote sinks are optional. The OpenAI Responses API, host tools, and provider invoices are unsupported. See [enforcement-boundary.md](https://github.com/bmdhodl/agent47/blob/main/docs/enforcement-boundary.md).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.