Skip to main content

PlanSpec + lint + safety gates for agent plans.

Project description

sdf-plan

Tool safety gates for agent workflows.

30-Second Quickstart (ToolGate-first)

from sdf_plan import confirm, propose

first = propose(
    tool_name="filesystem.write",
    args={"path": "/tmp/demo.txt", "content": "hello"},
    meta={"workspace_id": "demo-ws"},
    run_context={"workspace_id": "demo-ws"},
)
print(first.decision.value)  # BLOCK

token = first.resume.token
_ = confirm(token, user_ok=True)

second = propose(
    tool_name="filesystem.write",
    args={"path": "/tmp/demo.txt", "content": "hello"},
    meta={"workspace_id": "demo-ws", "confirmed_token": token},
    run_context={"workspace_id": "demo-ws"},
)
print(second.decision.value)  # ALLOW

Expected flow: BLOCKED -> CONFIRM -> CONTINUE

Install

pip install sdf-plan

5-Minute First Success

python examples/tool_gate_quickstart.py
python examples/tool_gate_openai_input.py
python examples/plan_mode_preflight.py

What You Get

  • ToolGate runtime decisions (ALLOW | WARN | BLOCK)
  • Signed confirmation tokens + resume flow
  • Idempotency key derivation from scope + tool + canonical args
  • Tool-mode lint rules + policy defaults
  • PlanSpec lint and preflight (optional mode)
  • LangGraph adapter (official thin wrapper)

Optional PlanSpec Mode

Plan mode remains supported for existing users.

from sdf_plan import lint_plan, policy_annotate, preflight_lint

plan = {
    "steps": [
        {
            "id": "S1",
            "type": "ACT",
            "title": "send email",
            "intent": "send email",
            "inputs": [],
            "outputs": ["ctx.sent"],
            "depends_on": [],
            "stop_condition": "Step S1 completed",
            "fallback": "reduce_scope",
            "idempotency_key": "idem-1",
        }
    ]
}
plan, summary = policy_annotate(plan)
findings = lint_plan(plan, max_steps=12, safety_mode="safe")
preflight_lint(plan, max_steps=12, safety_mode="safe")

Guides

  • docs/ADAPTER_TEMPLATE.md
  • docs/POLICY_TUNING.md
  • docs/TOOL_CLASSIFICATION.md
  • docs/COMPATIBILITY.md

Examples

  • examples/tool_gate_quickstart.py
  • examples/tool_gate_openai_input.py
  • examples/plan_mode_preflight.py
  • examples/adapter_minimal.py
  • examples/langgraph_plangate_demo.py
  • examples/crewai_plangate_demo.py

Compatibility

Use Cloud schema hash checks to detect contract drift:

from sdf_plan.compat import assert_schema_compat, package_version

assert_schema_compat(package_version(), "schema_hash_from_/v1/schema")

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

sdf_plan-0.2.5-py3-none-any.whl (28.8 kB view details)

Uploaded Python 3

File details

Details for the file sdf_plan-0.2.5-py3-none-any.whl.

File metadata

  • Download URL: sdf_plan-0.2.5-py3-none-any.whl
  • Upload date:
  • Size: 28.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for sdf_plan-0.2.5-py3-none-any.whl
Algorithm Hash digest
SHA256 94e09ad7cbdd2f0ba12ce67ceeb119b19a9cb53613b298d1cc40e79e9388e089
MD5 a22fdd3fc58f272e5088ac69fbf12ce0
BLAKE2b-256 9210c689177c29cd3bfa8a22a02a6217fdda6fbc480cc9c0b9fe10b8fd1bd8c1

See more details on using hashes here.

Provenance

The following attestation bundles were made for sdf_plan-0.2.5-py3-none-any.whl:

Publisher: release.yml on directiveproto/sdf-plan

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page