Files
composiohq__composio/python
Alberto Schiabel 6ba9179b48 fix(files): validate URLs from API responses before fetching them (#4146)
This PR:

- builds on top of https://github.com/ComposioHQ/composio/pull/4144
- routes every fetch whose URL comes from an API response through the
SSRF guards that already existed — ten sinks, five per SDK: the
tool-execution download, both S3 presigned uploads,
`RemoteFile.buffer()`/`blob()`, and the Tool Router session file upload
- adds `safe_request()` (Python), which follows redirects itself and
re-validates each hop, so a validated URL cannot 302 into private space
and an S3 307 region redirect still works
- makes `RemoteFile.buffer()` (Python) share `_fetch_url_bytes` with the
user-supplied-URL path instead of duplicating it — it previously read
`response.content` with no size cap, no redirect control, and no target
validation
- adds `ssrfSafeFetchWhereSupported` (TypeScript): the full guard on
Node, a plain `fetch` on workerd, so Tool Router session file transfers
keep working in edge runtimes rather than failing closed
- documents DNS rebinding as a known residual in both guards
- adds tests that assert the guard *runs* — blocked URL, sink never
called, nothing written — rather than that a transfer succeeds

## Context

`python/AGENTS.md` states the trust boundary: every field of an API
response is untrusted input, because the backend may be compromised or
the connection MITM'd. Both SDKs enforced that for URLs a *user* passes
in (`composio.utils.url_safety`, `ssrfGuard.node.ts`) and left the URLs
an API *response* supplies unguarded — backwards relative to the stated
model. A response naming an internal address turned the SDK into a
request proxy for it, and the fetched bytes were written to disk or
returned to the caller, typically into an LLM context.

`RemoteFile.buffer()` was the weakest of the ten: four lines above it,
`_fetch_from_url` validated its target, refused redirects, and streamed
against a 100 MiB cap; `buffer()` did none of the three. The only
difference between them was which side of the trust boundary the URL
arrived from.

Three decisions worth review:

- **The guard is unconditional.** Presigned URLs are public, so nothing
legitimate should resolve to private space. There is no escape hatch,
and no new configuration surface.
- **The tool-execution download stays uncapped.** It streams straight to
disk, so `_MAX_RESPONSE_SIZE` — a memory-exhaustion bound — does not
apply, and tool attachments legitimately exceed it.
`RemoteFile.buffer()` *is* capped, because it buffers in memory.
- **Uploads follow redirects with re-validation rather than refusing
them**, since S3 can answer a PUT with a 307 region redirect. Downloads
keep `allow_redirects=False`, matching the existing fetch paths.

Python now matches the TypeScript guard's per-hop re-validation, which
was the better of the two implementations.

Verified locally: `make chk` and `make tst` clean on the Python side;
`pnpm -C packages/core test` 1095 passed, typecheck and lint clean.
2026-08-18 13:25:01 +02:00
..
2025-06-17 12:58:00 +05:30
2025-06-17 12:58:00 +05:30

Composio logo

Composio Python SDK

Composio gives your AI agents 1000+ pre-authenticated toolkits, per-user sessions, authentication, triggers, and a sandbox. This package is the Python SDK.

Install

Requires Python 3.10+.

pip install composio

Quickstart

Create a session for one of your users and hand its tools to your agent:

from composio import Composio

composio = Composio()  # reads COMPOSIO_API_KEY, or pass api_key=...

session = composio.create(user_id="user_123")
tools = session.tools()  # OpenAI-format tool definitions by default

By default a session gets meta tools that discover, authenticate, and execute app tools at runtime, so you don't load hundreds of tool definitions into context. Store session.session_id and reuse the session across turns:

session = composio.use(session_id)

See how Composio works for sessions and meta tools, and configuring sessions for restricting toolkits, tools, auth_configs, and connected_accounts on composio.create().

Use with an agent framework

Provider packages adapt session.tools() to your framework's native tool format. With OpenAI Agents:

pip install composio composio-openai-agents openai-agents
from composio import Composio
from composio_openai_agents import OpenAIAgentsProvider
from agents import Agent, Runner

composio = Composio(provider=OpenAIAgentsProvider())

session = composio.create(user_id="user_123")
tools = session.tools()

agent = Agent(
    name="Personal Assistant",
    instructions="You are a helpful assistant. Use Composio tools to take action.",
    tools=tools,
)

result = Runner.run_sync(starting_agent=agent, input="Summarize my emails from today")
print(result.final_output)

Other Python providers: composio-openai, composio-anthropic, composio-claude-agent-sdk, composio-langchain, composio-langgraph, composio-llamaindex, composio-crewai, composio-autogen, composio-gemini, composio-google, composio-google-adk. Don't see yours? Build a custom provider.

MCP

Every session also exposes a hosted MCP endpoint. Pass mcp=True and point Claude, Cursor, or any MCP client at it:

from composio import Composio

composio = Composio()

session = composio.create(user_id="user_123", mcp=True)
print(session.mcp.url)      # MCP endpoint for this session
print(session.mcp.headers)  # auth headers for the endpoint

See sessions via MCP.

Authentication

Sessions manage connections for you by default; the agent walks the user through OAuth with the session's meta tools. To drive the flow yourself, authorize a toolkit from the session:

connection_request = session.authorize("gmail")
print(connection_request.redirect_url)  # send the user here to approve access
connection_request.wait_for_connection()

See authentication for auth configs, custom OAuth apps, and connection lifecycle.

Triggers

Subscribe to events from connected apps (new email, new commit, and so on) and react to them:

from composio import Composio

composio = Composio()

trigger = composio.triggers.create(
    slug="GITHUB_COMMIT_EVENT",
    user_id="user_123",
    trigger_config={"owner": "composiohq", "repo": "composio"},
)

subscription = composio.triggers.subscribe()

@subscription.handle(trigger_id=trigger.trigger_id)
def handle_event(data):
    print("Event received:", data)

subscription.wait_forever()

subscribe() streams events over a WebSocket for local development. In production, register a webhook URL and parse deliveries with composio.triggers.parse(). See setting up triggers.

Development

This package lives in the Composio SDK monorepo under python/. See the contribution guidelines to get set up.

Support