agentic-sidecar

A companion intelligence and real-time supervision layer for autonomous AI agents.

The Main Agent acts. The Sidecar observes, thinks, advises, and governs.

Status

v0.1 and v0.2 implemented. The Sidecar runtime, a rule-based Decision Gate (Policy Advisor + Risk Evaluator), and Intent Guardian (IntentEnvelope

• constraint validation) are real code, attached via the LangGraph adapter. Observe mode (v0.1, logs only) and Govern mode (v0.2, a BLOCK is actually enforced) both work today — see Python API (implemented) below and examples/ for two runnable scripts. No PyPI release yet. See ROADMAP.md for the full build plan (v0.2.x onward).

Contents

• Why
• What it is not
• Architecture
• Sidecar vs. Agent Harness
• Main Agent vs. Sidecar
• How the Decision Gate evaluates a decision
• Sidecar modules
• Python API (implemented)
• Planned Python API
• Operating modes
• Long-term: an intent propagation layer
• The DeepAgentLabs ecosystem
• Roadmap
• License

Why

Autonomous agents chain a lot of reasoning and tool calls between a user's request and a real-world effect:

User → Agent → Planning → Sub-Agent → MCP → Tool → API → External System

Three things go wrong in that chain that permission checks alone don't catch:

• Intent drift — "investigate why production is slow" quietly becomes "restart the production database." Nobody revoked authority; the agent's plan just wandered.
• Technically-allowed, contextually-wrong actions — refund_customer() is a permitted tool call. Refunding $850 against a $500 authorization is not a permitted decision, even though the tool call succeeds.
• No live visibility — long-running agents act for minutes at a time with no answer to "what is it doing right now, and can I stop it?"

agentic-sidecar is designed to attach to any agent framework, but the first release proves that against one framework only — LangGraph — before claiming the rest (see Design Constraints in ROADMAP.md). Either way, it answers a different question than permission checks or observability tools do:

Not "can this agent call this tool?" — "should it, right now, given what the user actually asked for?"

What it is not

• Not a multi-agent framework. The Sidecar never owns or executes the task plan, and it never replaces a worker agent. It may independently critique the Main Agent's plan or propose alternatives (the Planner module) — but the Main Agent decides whether to act on that critique, and remains the one that executes.
• Not a content-safety guardrail. Libraries like NeMo Guardrails or Guardrails AI validate a single prompt/response turn against rules. The Sidecar evaluates a decision — a tool call or plan step — against intent and context accumulated across an entire task.
• Not a tracing/observability tool. That's AgenticLens's job: "what did the agent do, and why did it fail?" — after the fact. The Sidecar's job is "should it continue?" — while it's still running.
• Not just LLM-as-a-Judge. A judge model is one possible evaluator inside one Sidecar module (§ Sidecar modules). Most decisions should never reach an LLM at all — see Operating modes and the cost-control design in ROADMAP.md.
• Not an agent harness. It doesn't run the agent loop, manage tools, or handle retries — that's LangGraph's, CrewAI's, or your custom loop's job. The Sidecar attaches to whatever harness is already running the agent; see Sidecar vs. Agent Harness.

Architecture

                    USER
                      │
                 User Intent
                      │
                      ▼
              ┌───────────────┐
              │  MAIN AGENT   │
              │   EXECUTOR    │
              └───────┬───────┘
                      │
              Plans / Decisions
                      │
                      ▼
            ┌─────────────────────┐
            │   AGENTIC SIDECAR   │
            │                     │
            │ Intent Guardian     │
            │ Planner             │
            │ Critic              │
            │ Judge               │
            │ Risk Evaluator      │
            │ Policy Advisor      │
            │ Decision Gate       │
            │ Status Interpreter  │
            └──────────┬──────────┘
                       │
          Advice / Approval / Challenge
                       │
                       ▼
                 MAIN AGENT
                       │
                       ▼
                  MCP / Tools
                       │
                       ▼
                External Systems

The Sidecar is attached to the execution lifecycle. It does not own the user's task and is never a second worker in the plan.

Sidecar vs. Agent Harness

An agent harness (LangGraph, a custom loop, OpenAI Agents SDK, Microsoft Agent Framework, ...) is the control and execution environment: it runs the agent loop and manages tools, state, context, retries, and lifecycle. agentic-sidecar doesn't replace that — it attaches to it as a decision-time supervision layer:

Agent Harness   = control and execution loop.
Agentic Sidecar = decision-time supervision layer for that loop.

The harness answers "how do I execute this workflow?" The Sidecar answers "should this decision happen, given what the human originally asked for?"

This distinction matters most once a task fans out across a delegation chain — Agent A → Agent B → Agent C → MCP/Tool — where each hop tends to receive only the sub-task it needs to perform, not the original constraints and authority behind it:

Human Intent → Agent A → delegates → Agent B → delegates → Agent C → MCP/Tool

The Sidecar's Intent Envelope (§ Sidecar modules) is designed to travel with the task through that chain instead of being reconstructed from conversation history at every hop — see Long-term: an intent propagation layer.

Where it pays for itself: without a supervision layer, a harness typically discovers a bad decision only after acting on it —

Plan → Act → Fail → Recover → Replan

— versus catching it before execution:

Plan → Sidecar Check → Act
          ├── ALLOW
          ├── CHALLENGE
          ├── REPLAN
          ├── BLOCK
          └── ESCALATE

That's primarily a reliability and control win, not a speed win — the Sidecar doesn't make the underlying model faster, it reduces unnecessary actions, unsafe retries, and repeated context reconstruction. Full treatment: concept.md § 23.

Main Agent vs. Sidecar

How the Decision Gate evaluates a decision

Three modules ask three genuinely different questions about the same proposed action, and the roadmap's job is to keep them from collapsing into one generic "policy engine":

             USER INTENT
                  │
          Intent Envelope
                  │
                  ▼
             MAIN AGENT
                  │
           proposed action
                  │
                  ▼
              SIDECAR
          ┌───────┼────────┐
          │       │        │
       Policy    Risk    Intent
          │       │        │
   "Are you    "How      "Is this
   permitted   dangerous  actually what
   to do       is this    the human
   this?"      action?"   asked you to
                          accomplish?"
          │       │        │
          └───────┼────────┘
                  │
             DECISION GATE
                  │
       ┌──────────┼──────────┐
       ▼          ▼          ▼
     ALLOW      REPLAN     ESCALATE
                              │
                            HUMAN

Policy and Risk are largely mechanical — permission lists, thresholds, tool-argument patterns — and v0.1 ships them with zero LLM calls (see ROADMAP.md). Intent (v0.2) is still deterministic, not LLM-based, but is the one that gets closest to "what the user actually meant" so far, via structured IntentEnvelope constraints rather than free-form judgment — full semantic understanding is further out on the roadmap.

The diagram above shows the target outcome set. v0.1 shipped a narrower slice (ALLOW/BLOCK only, Observe mode: the Sidecar logs what it would have decided but cannot stop an action in practice). v0.2 adds WARN (an Intent Guardian finding worth surfacing but not severe enough to block, e.g. a stale envelope) and Govern mode, where a BLOCK is actually enforced by the attached adapter — see agentic_sidecar.core.exceptions.SidecarBlockedError. CHALLENGE, REPLAN, PAUSE, and ESCALATE remain v0.4. Version-by-version build order is in ROADMAP.md.

Sidecar modules

Users enable only what they need:

Agentic Sidecar
│
├── Intent Guardian     — Intent Envelope, alignment checks, drift detection
├── Planner              — independently evaluates the agent's plan
├── Critic                — challenges a proposed decision before it executes
├── Judge                 — optional independent (and independent-model) evaluation
├── Risk Evaluator        — classifies an action's risk before deciding whether to escalate
├── Policy Advisor        — deterministic policy rules (cheapest check, runs first)
├── Decision Gate         — turns evaluations into ALLOW / WARN / CHALLENGE / REPLAN / PAUSE / BLOCK / ESCALATE
├── Budget Guardian       — cost/token ceilings per task
├── Status Interpreter    — translates raw tool/MCP traces into human-readable narration
└── Human Escalation      — pauses execution and requests approval

sidecar:
  intent:
    enabled: true
    preserve_original_intent: true
  planner:
    enabled: false        # off by default — see cost design in ROADMAP.md
  critic:
    enabled: false        # off by default — see cost design in ROADMAP.md
  policy:
    enabled: true
    source: policies.yaml
  risk:
    enabled: true
    intervention_threshold: 0.80
  judge:
    enabled: false        # off by default — see cost design in ROADMAP.md
    model: independent-model
  budget:
    enabled: true
    max_cost: 2.00
  human_approval:
    enabled: true
  on_sidecar_failure: fail_closed   # or fail_open — see ROADMAP.md

Python API (implemented)

v0.1 (Policy Advisor + Risk Evaluator, Observe mode):

from agentic_sidecar import Sidecar
from agentic_sidecar.adapters.langgraph import attach
from agentic_sidecar.gate.policy import PolicyAdvisor, PolicyRule
from agentic_sidecar.gate.risk import RiskEvaluator, RiskRule

sidecar = Sidecar(
    on_sidecar_failure="fail_closed",  # required, no default -- see Design Constraints
    policy=PolicyAdvisor([PolicyRule(tool="delete_*", effect="deny")]),
    risk=RiskEvaluator(
        [RiskRule(tool="issue_refund", arg_name="amount", arg_op="gt", arg_value=500, risk="HIGH")]
    ),
)

# Wrap the tools you'd otherwise pass straight to
# langgraph.prebuilt.create_react_agent(model, tools=...) -- every call is
# evaluated by the Sidecar first.
tools = attach(sidecar, [read_order, issue_refund, delete_customer_record])
agent = create_react_agent(model, tools=tools)

# mode="observe" is the default: agent.invoke(...) still executes every
# tool call. Nothing is enforced -- inspect what the Sidecar would have
# decided:
for context, decision in sidecar.decisions:
    print(decision.status, decision.risk, context.tool_name, decision.reason)

v0.2 adds Intent Guardian and Govern mode, where BLOCK is actually enforced -- the refund-limit scenario from concept.md §9:

from agentic_sidecar import Sidecar, SidecarBlockedError
from agentic_sidecar.adapters.langgraph import attach
from agentic_sidecar.intent.alignment import ConstraintBinding, IntentGuardian
from agentic_sidecar.intent.envelope import IntentEnvelope, Requester

envelope = IntentEnvelope(
    goal="refund_customer",
    requested_by=Requester(type="human", id="user123"),
    constraints={"maximum_refund": 500},
)
binding = ConstraintBinding(
    constraint="maximum_refund", tool="issue_refund", arg_name="amount", op="lte"
)

sidecar = Sidecar(
    on_sidecar_failure="fail_closed",
    roles=["policy", "risk", "intent_guardian"],
    intent=IntentGuardian(envelope, [binding]),
    mode="govern",  # a BLOCK now actually stops the call
)
tools = attach(sidecar, [issue_refund])
agent = create_react_agent(model, tools=tools)

try:
    agent.invoke({"messages": [("user", "Refund order A100 for $850.")]})
except SidecarBlockedError as exc:
    print(exc)  # "Sidecar blocked 'issue_refund'(...): Intent Guardian: ..."

sidecar.set_intent(guardian) swaps the active envelope between tasks -- IntentEnvelopes are meant to be per-task (concept.md §6), not fixed for a Sidecar's whole lifetime.

Sidecar.before_tool_call registers a custom hook that fully replaces the default Policy + Risk + Intent Guardian evaluation:

@sidecar.before_tool_call
def evaluate_action(context):
    if context.tool_name == "delete_customer_record":
        return Decision(status="BLOCK", risk="HIGH", reason="never allowed")
    return Decision(status="ALLOW", risk="LOW", reason="ok")

Note the shape: agentic_sidecar.adapters.langgraph.attach(sidecar, tools), not sidecar.attach(my_agent). core/ intentionally has zero dependency on adapters/ (see AGENTS.md's Package Boundaries), so wrapping a specific framework's decision boundaries is the adapter's job, not a generic method on Sidecar. The "Planned Python API" below is the longer-term target this is expected to grow toward once framework independence has actually been proven (ROADMAP.md's Design Constraint 1) — not a regression.

Planned Python API

This is the longer-term target developer experience once a second framework adapter exists and roles covers Planner/Critic too (v0.3) — not yet implemented. roles=["intent_guardian", "policy", "risk"] and Sidecar(...) construction already work today (see Python API (implemented) above); what's still aspirational here is specifically sidecar.attach(my_agent) wrapping the whole agent generically (today it's per-adapter, e.g. agentic_sidecar.adapters.langgraph.attach) and sidecar.evaluate(intent=, action=, risk=)'s keyword-argument call shape:

from agentic_sidecar import Sidecar

sidecar = Sidecar(roles=["intent_guardian", "policy", "risk", "planner", "critic"])
agent = sidecar.attach(my_agent)

agent.run("Investigate the production issue but do not modify production.")

@sidecar.before_tool_call
def evaluate_action(context):
    return sidecar.evaluate(
        intent=context.intent,
        action=context.tool_call,
        risk=context.risk,
    )

Decision(
    status="REPLAN",
    risk="HIGH",
    reason="Action exceeds original user intent",
)

Operating modes

Modes are meant to be adopted in that order — organizations start in Observe and move to Govern once they trust the signal.

Long-term: an intent propagation layer

The nearer-term modules above are the whole of what v0.1–v0.8 ship. But the IntentEnvelope (§ Sidecar modules) is designed to survive being handed off — not just checked once and discarded:

Human
  │
  └── Intent Envelope #182
            │
            ▼
         Agent A
            │
        delegates
            ▼
         Agent B
            │
           MCP
            ▼
         Agent C
            │
           Tool

If every hop in a delegation chain can answer what was originally requested, who authorized it, what constraints apply, what authority was actually delegated, and whether intent has since changed — that's no longer just a feature of one package. It's closer to an interoperability concern for autonomous systems generally, which is why v1.0 targets publishing the envelope as a versioned schema in ai-operations-spec rather than keeping it as an internal Sidecar structure (see the v1.0 entry in ROADMAP.md). This is explicitly a v1.0 target, not something v0.1 needs to anticipate — noted here because it's the reason the envelope's shape deserves care early, even though nothing consumes it across process boundaries yet.

The DeepAgentLabs ecosystem

                 DeepAgentLabs
              Autonomous AI Systems
                      │
        ┌─────────────┼─────────────┐
        │             │             │
     OBSERVE         GUIDE        CONNECT
        │             │             │
 AgenticLens    Agentic Sidecar  Agentic MCP
        │             │             │
        └─────────────┼─────────────┘
                      │
                     TEST
                      │
                Agentic Chaos

Each project is independently installable; none requires another as a hard dependency (see Cross-Project Dependencies in the roadmap).

Roadmap

Full build plan, version sequencing, and design constraints: ROADMAP.md.

Original architecture proposal: concept.md.

License

MIT (planned — LICENSE file to be added alongside the first code commit, matching sibling DeepAgentLabs projects).