agent-failsafe
FailSafe governance adapter for the Microsoft Agent Governance Toolkit.
agent-failsafe is a Python adapter that connects the FailSafe runtime governance engine to the Microsoft Agent Governance Toolkit.
• Adds runtime behavioral governance to the toolkit
• Detects failure patterns such as tool recursion and reasoning collapse
• Integrates through interceptors, validators, and policy providers
• Install with pip install agent-failsafe
Core runtime: https://github.com/MythologIQ/Failsafe
Installation
# Core only (types, clients, patterns, trust scoring)
pip install agent-failsafe
# With specific toolkit packages
pip install agent-failsafe[agent-os] # Interceptor + Integration + Webhook events
pip install agent-failsafe[agent-sre] # SLI + Signal generation
pip install agent-failsafe[agent-mesh] # Audit sink
pip install agent-failsafe[agent-hypervisor] # ExecutionRing + KillSwitch
# Everything
pip install agent-failsafe[full]
The public extras above keep their names, but resolve to the consolidated
Microsoft Agent Governance Toolkit 4.x packages: agent-os, agent-mesh, and
agent-hypervisor pull agent-governance-toolkit-core (>=4.0.0,<5.0), while
agent-sre pulls agent-governance-toolkit-cli (>=4.0.0,<5.0). full (and
dev) install the union of both consolidated packages.
Extension Points
| Adapter | Toolkit Extension Point | Module |
|---|---|---|
FailSafeInterceptor |
ToolCallInterceptor |
interceptor.py |
FailSafeKernel |
BaseIntegration + @register_adapter |
integration.py |
GovernancePipeline |
Full lifecycle orchestration | pipeline.py |
FailSafeRingAdapter |
ExecutionRing + KillSwitch |
ring_adapter.py |
FailSafeTrustValidator |
ValidatorInterface (control plane) |
trust_validator.py |
ShadowGenomePolicyProvider |
PolicyProviderInterface |
policy_provider.py |
FailSafeComplianceSLI |
SLI (agent-sre) |
sli.py |
FailSafeAuditSink |
AuditSink |
audit_sink.py |
FailSafeApprovalBackend |
ApprovalBackend |
escalation.py |
FailSafeTrustMapper |
DID translation (did:myth <> did:mesh) |
trust_mapper.py |
decision_to_webhook_event |
WebhookEvent translation |
webhook_events.py |
Architecture Diagram
AI Agent
↓
Agent Governance Toolkit
↓
agent-failsafe adapter
↓
FailSafe runtime engine
↓
Tool execution
Basic Usage Example
Basic interceptor example
from agent_failsafe import FailSafeInterceptor
interceptor = FailSafeInterceptor()
result = interceptor.intercept(request)
from agent_failsafe import FailSafeInterceptor
interceptor = FailSafeInterceptor()
result = interceptor.intercept(request)
See docs/ADAPTER_ARCHITECTURE.md for full API reference.
Ecosystem
FailSafe runtime engine https://github.com/MythologIQ/Failsafe
Developer extension https://marketplace.visualstudio.com/items?itemName=MythologIQ.mythologiq-failsafe https://open-vsx.org/extension/MythologIQ/mythologiq-failsafe
Recommended GitHub Topics
- ai-agents
- agent-governance
- agentic-ai
- ai-runtime
- ai-reliability
- llm-agents
- ai-infrastructure
- open-source
License
MIT
Metadata
Release files for agent-failsafe 0.6.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agent_failsafe-0.6.1.tar.gz | 99.2 kB | With provenance | Details for agent_failsafe-0.6.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agent_failsafe-0.6.1-py3-none-any.whl | Python 3 | none | any | Details for agent_failsafe-0.6.1-py3-none-any.whl |
Total release size: 148.2 kB
Release files / agent_failsafe-0.6.1.tar.gz
| Download URL | agent_failsafe-0.6.1.tar.gz |
|---|---|
| Size | 99.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
bf6261a8581c6c13b67a6dd3707bde4a1a88f058abfff7b166814f505b2f0c82
|
|
BLAKE2b-256 checksum How to use checksums |
41cc3bbc3dc9811c15329759eb50d96658e0140c5de0fa5e9168013ac2ae8e62
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jun 3, 2026.
Transparency logRelease files / agent_failsafe-0.6.1-py3-none-any.whl
| Download URL | agent_failsafe-0.6.1-py3-none-any.whl |
|---|---|
| Size | 49.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
7c0337c9322ff5b9639b464ea428b8eab5cb5f01676613a5ef4c8101bd08d38f
|
|
BLAKE2b-256 checksum How to use checksums |
5bdc36eb6b761c444c48a28ec9fdd2eebce70a5733fb363d8d24f6898bb8a2d7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jun 3, 2026.
Transparency logSecurity for agent-failsafe 0.6.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for agent_failsafe-0.6.1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jun 3, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.