headroom/tests/test_auth_mode.py
Abhay Singh 8cddf9b58e
fix(proxy/auth): match real Anthropic OAuth token prefix (sk-ant-oat) (#1672)
## Description

`classify_auth_mode` (in `headroom/proxy/auth_mode.py`) checks for
Anthropic
OAuth tokens with:

```python
if token.startswith("sk-ant-oat-"):
    return AuthMode.OAUTH
if token.startswith("sk-ant-api") or token.startswith("sk-"):
    return AuthMode.PAYG
```

But real Anthropic OAuth access tokens are **`sk-ant-oat01-...`** — a
version
number right after `oat`, **no dash**. So the `sk-ant-oat-` check never
matches a
real token; it falls through to the broad `sk-` rule and gets classified
**`PAYG`**.

That's exactly the misclassification the module is built to prevent: a
subscription/OAuth-bound request tagged `PAYG` gets the
aggressive-compression
policy — lossy compression, auto `cache_control`, `prompt_cache_key`
injection —
instead of the passthrough-prefer path OAuth is meant to get.

The existing tests didn't catch it because they use a synthetic
`sk-ant-oat-01-`
fixture (dashed) that happens to match the buggy prefix. Corroboration
that the
real shape is dash-less:
- `.gitguardian.yaml` fixture: `sk-ant-oat01-oauth-fixture`
- `tests/test_oauth_bearer_routing.py`: `sk-ant-oat01-xxx`
- the sibling helper `headroom/proxy/helpers.py` matches on `sk-ant-`
(no `oat-`)

## Fix

Match the dash-less `sk-ant-oat` prefix. It still matches the legacy
dashed
shape, and ordering relative to `sk-ant-api` / `sk-` is unchanged (OAuth
is
still checked first).

```python
if token.startswith("sk-ant-oat"):
    return AuthMode.OAUTH
```

## Type of Change

- [x] Bug fix (non-breaking change that fixes an issue)

## Changes Made

- `headroom/proxy/auth_mode.py`: match OAuth tokens on the dash-less
`sk-ant-oat` prefix.
- `tests/test_auth_mode.py`: add a regression test using the real
`sk-ant-oat01-...` format (the existing test keeps the legacy dashed
fixture, which still classifies correctly).
- `CHANGELOG.md`: Bug Fixes entry under Unreleased.

## Testing

- [x] New regression test added (`tests/test_auth_mode.py`)
- [x] Linting passes (`ruff check`) and formatting is clean (`ruff
format --check`)
- [ ] Full `pytest` deferred to CI (local-OOM reason below).

```text
$ uv run ruff check headroom/proxy/auth_mode.py tests/test_auth_mode.py
All checks passed!
```

## Real Behavior Proof

- Environment: Windows 11, Python 3.12.11, headroom from this branch.
Importing `headroom` loads the torch/transformers stack and a full
`pytest` gets OOM-killed on this box, so I verified the classification
logic with a dependency-free script and left the full pytest to CI.
- Exact command / steps: replicated the Bearer-token branch of
`classify_auth_mode` in a standalone script (only stdlib, no `headroom`
import) and ran the real and legacy token shapes plus PAYG keys through
it.
- Observed result: the real `sk-ant-oat01-...` now classifies OAUTH (was
PAYG before the change); the legacy dashed fixture still classifies
OAUTH; `sk-ant-api*` / `sk-*` keys still classify PAYG:

```text
OK: sk-ant-oat01-... -> OAUTH (was PAYG before fix)
OK: sk-ant-oat-01-... -> OAUTH (legacy fixture still matches)
OK: sk-ant-api* / sk-* -> PAYG (unchanged)
AUTH LOGIC VERIFIED
```

- Not tested: a live proxied Anthropic OAuth request end-to-end (needs a
real subscription token); the classification is pure and covered by the
regression test. Full local `pytest` deferred to CI (OOM, per above).

## Review Readiness

- [x] I have performed a self-review
- [x] This PR is ready for human review

## Checklist

- [x] My code follows the project's style guidelines
- [x] I have performed a self-review of my code
- [x] I have commented my code, particularly in hard-to-understand areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes — ran
lint + a standalone logic check; full pytest deferred to CI (local OOM,
disclosed above)
- [x] I have updated the CHANGELOG.md if applicable

## Additional Notes

- `crates/headroom-core/src/auth_mode.rs` carries the identical dashed
prefix (its Rust test matrix uses the same synthetic dashed fixture). I
scoped this PR to the Python runtime classifier since that's the
request-time path; happy to mirror the one-line fix in Rust in the same
PR or a follow-up — I just couldn't `cargo build` locally to verify, so
I left it out rather than push an unverified Rust edit.
- @JerrettDavis tagging you since you've been triaging these — small,
contained fix with a regression test if you have a moment.

Co-authored-by: JerrettDavis <mxjerrett@gmail.com>
2026-07-02 14:26:19 -07:00

204 lines
7.5 KiB
Python

"""Python parity tests for ``headroom.proxy.auth_mode.classify_auth_mode``.
Mirrors the Rust matrix in ``crates/headroom-core/tests/auth_mode.rs``
byte-for-byte. The two implementations MUST agree on every header set
covered here. Any divergence is a bug — we catch it at PR review by
running both suites side by side.
"""
from __future__ import annotations
import time
import pytest
from headroom.proxy.auth_mode import (
SUBSCRIPTION_UA_PREFIXES,
AuthMode,
classify_auth_mode,
classify_client,
)
def _h(**pairs: str) -> dict[str, str]:
"""Build a plain-dict header set in one expression."""
return dict(pairs)
# ── Required matrix ──────────────────────────────────────────────
def test_api_key_classified_payg() -> None:
"""Anthropic PAYG: ``Authorization: Bearer sk-ant-api03-...``."""
headers = {"authorization": "Bearer sk-ant-api03-abc123def456"}
assert classify_auth_mode(headers) is AuthMode.PAYG
def test_oauth_jwt_classified_oauth() -> None:
"""Codex / Cursor OAuth bearer: classic 3-segment JWT."""
jwt = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0In0.signaturepart"
headers = {"authorization": f"Bearer {jwt}"}
assert classify_auth_mode(headers) is AuthMode.OAUTH
def test_oauth_sk_ant_oat_classified_oauth() -> None:
"""Claude Pro / Max OAuth: ``Bearer sk-ant-oat-...``."""
headers = {"authorization": "Bearer sk-ant-oat-01-abc123def456"}
assert classify_auth_mode(headers) is AuthMode.OAUTH
def test_oauth_real_sk_ant_oat01_classified_oauth() -> None:
"""Real Anthropic OAuth access tokens are ``sk-ant-oat01-...`` (a version
number, no dash after ``oat``). These must classify as OAUTH — matching on
``sk-ant-oat-`` missed them and let them fall through to PAYG, enabling
aggressive lossy compression on subscription-bound requests."""
headers = {"authorization": "Bearer sk-ant-oat01-abc123def456"}
assert classify_auth_mode(headers) is AuthMode.OAUTH
def test_claude_code_ua_classified_subscription() -> None:
"""Claude Code CLI: ``User-Agent: claude-code/1.2.3 ...``."""
headers = {"user-agent": "claude-code/1.2.3 (darwin; arm64)"}
assert classify_auth_mode(headers) is AuthMode.SUBSCRIPTION
def test_cursor_ua_classified_subscription() -> None:
"""Cursor CLI: ``User-Agent: cursor/1.0``."""
headers = {"user-agent": "cursor/1.0"}
assert classify_auth_mode(headers) is AuthMode.SUBSCRIPTION
def test_no_auth_no_user_agent_default_payg() -> None:
"""Empty headers → safest default is PAYG.
The OAuth/bedrock branch fires only when there's a positive
non-Bearer auth signal. Choosing PAYG by default favors the
OSS-default workload (per-token cost saving).
"""
assert classify_auth_mode({}) is AuthMode.PAYG
def test_bedrock_no_auth_classified_oauth() -> None:
"""Bedrock SigV4: ``Authorization: AWS4-HMAC-SHA256 Credential=...``.
Not a Bearer scheme; we treat all non-Bearer Authorization as
OAuth (passthrough-prefer).
"""
headers = {
"authorization": (
"AWS4-HMAC-SHA256 Credential=AKIAIOSFODNN7EXAMPLE/20260501/"
"us-east-1/bedrock/aws4_request, SignedHeaders=host;x-amz-date, "
"Signature=fe5f80f77d5fa3beca038a248ff027"
),
}
assert classify_auth_mode(headers) is AuthMode.OAUTH
# ── Bonus matrix ──────────────────────────────────────────────────
def test_openai_payg_sk_classified_payg() -> None:
"""OpenAI PAYG: ``Authorization: Bearer sk-proj-...``."""
headers = {"authorization": "Bearer sk-proj-abcdef0123456789"}
assert classify_auth_mode(headers) is AuthMode.PAYG
def test_gemini_x_goog_api_key_classified_payg() -> None:
"""Google Gemini API key as ``x-goog-api-key``."""
headers = {"x-goog-api-key": "AIzaSyDUMMYKEY1234567890"}
assert classify_auth_mode(headers) is AuthMode.PAYG
def test_subscription_takes_precedence_over_oauth_token() -> None:
"""Claude Code CLI sends ``Bearer sk-ant-oat-...`` but is subscription.
UA wins because the rate-limit / fingerprint policy is bound to
the CLI, not the token shape. The same token via a non-CLI UA
would be OAuth.
"""
headers = {
"user-agent": "claude-code/1.5.0 (linux; x86_64)",
"authorization": "Bearer sk-ant-oat-01-abc123",
}
assert classify_auth_mode(headers) is AuthMode.SUBSCRIPTION
# ── Edge cases (defensive coverage) ──────────────────────────────
def test_anthropic_x_api_key_classified_payg() -> None:
"""Anthropic API key style: ``x-api-key: sk-ant-...``."""
headers = {"x-api-key": "sk-ant-api03-abcdef"}
assert classify_auth_mode(headers) is AuthMode.PAYG
@pytest.mark.parametrize("prefix", SUBSCRIPTION_UA_PREFIXES)
def test_every_subscription_prefix_classified_subscription(prefix: str) -> None:
"""Each entry in the prefix list classifies as Subscription on its own."""
headers = {"user-agent": f"{prefix}1.0 (test)"}
assert classify_auth_mode(headers) is AuthMode.SUBSCRIPTION
def test_unparseable_authorization_does_not_raise() -> None:
"""Bytes-valued non-UTF-8 authorization warns and falls through.
Mirrors the Rust ``classify`` behaviour: never panic, always
return a valid ``AuthMode``. With nothing else to disambiguate,
we land on PAYG (the safe default).
"""
headers = {"authorization": b"\xffnope"}
# Should not raise.
result = classify_auth_mode(headers)
assert result is AuthMode.PAYG
def test_case_insensitive_header_lookup() -> None:
"""Header names are matched case-insensitively (Starlette parity)."""
headers = {"Authorization": "Bearer sk-ant-api03-test"}
assert classify_auth_mode(headers) is AuthMode.PAYG
def test_enum_values_match_rust_as_str() -> None:
"""The string form of each enum member matches the Rust ``as_str()``."""
assert AuthMode.PAYG.value == "payg"
assert AuthMode.OAUTH.value == "oauth"
assert AuthMode.SUBSCRIPTION.value == "subscription"
# ── Performance ──────────────────────────────────────────────────
def test_classify_under_100us_per_call() -> None:
"""Smoke perf check.
Python is slower than Rust by an order of magnitude on string
work, so the budget here is 100us (vs 10us in Rust). The proxy
only calls this once per request, and the absolute ceiling is
well under 1ms; this guards against pathological regressions.
"""
headers = {
"user-agent": "claude-code/1.5.0 (linux; x86_64) anthropic/0.42.0",
"authorization": "Bearer sk-ant-oat-01-abcdefghijklmnopqrstuv",
"content-type": "application/json",
"accept": "application/json",
"host": "api.anthropic.com",
}
# Warmup
for _ in range(1000):
classify_auth_mode(headers)
iters = 10_000
start = time.perf_counter()
for _ in range(iters):
classify_auth_mode(headers)
elapsed = time.perf_counter() - start
per_call_us = (elapsed / iters) * 1_000_000
assert per_call_us < 100, f"classify_auth_mode took {per_call_us:.2f} us/call (limit: 100 us)"
def test_classify_client_uses_default_when_no_client_signal():
headers = {"user-agent": "anthropic/0.42.0"}
assert classify_client(headers, default="claude") == "claude"