Impact
aiograpi 0.6.6 / 0.7.0 / 0.7.1 declared orjson==3.11.6 (and later ==3.11.8) in requirements.txt but setup.py carried a hard-coded duplicate requirements = [...] list that was never updated and still pinned orjson==3.11.4.
When setuptools builds the source distribution it reads the metadata from setup.py, not from requirements.txt. So pip install aiograpi==0.6.6 (or 0.7.0 / 0.7.1) actually pulls orjson==3.11.4 — a version vulnerable to CVE-2025-67221 (stack overflow in orjson.dumps on deeply nested JSON inputs).
Practical exploitability
Low in the typical aiograpi flow: orjson is used to encode request bodies aiograpi itself constructs and to decode responses returned by Instagram. An attacker would need to coerce aiograpi to encode an attacker-controlled deeply-nested Python structure or to decode an attacker-supplied stream — not the normal call shape.
However any caller doing client.public_request(...) or similar with caller-controlled payloads, or any caller passing aiograpi-decoded last_json into recursive serialization, may hit the unbounded recursion. The patched orjson rejects deeply-nested inputs cleanly.
Patches
Fixed in aiograpi 0.7.2 by migrating to pyproject.toml (PEP 621) — single source of truth for dependencies. PyPI installs of 0.7.2 and later resolve orjson==3.11.8 correctly.
Workarounds
Force-install a non-vulnerable orjson alongside the affected aiograpi version:
pip install 'aiograpi==0.7.1' 'orjson>=3.11.6'
Or just upgrade to a fixed aiograpi:
pip install -U 'aiograpi>=0.7.2'
Resources
References
Impact
aiograpi 0.6.6 / 0.7.0 / 0.7.1 declared
orjson==3.11.6(and later==3.11.8) inrequirements.txtbutsetup.pycarried a hard-coded duplicaterequirements = [...]list that was never updated and still pinnedorjson==3.11.4.When
setuptoolsbuilds the source distribution it reads the metadata fromsetup.py, not fromrequirements.txt. Sopip install aiograpi==0.6.6(or 0.7.0 / 0.7.1) actually pullsorjson==3.11.4— a version vulnerable to CVE-2025-67221 (stack overflow inorjson.dumpson deeply nested JSON inputs).Practical exploitability
Low in the typical aiograpi flow:
orjsonis used to encode request bodies aiograpi itself constructs and to decode responses returned by Instagram. An attacker would need to coerce aiograpi to encode an attacker-controlled deeply-nested Python structure or to decode an attacker-supplied stream — not the normal call shape.However any caller doing
client.public_request(...)or similar with caller-controlled payloads, or any caller passing aiograpi-decodedlast_jsoninto recursive serialization, may hit the unbounded recursion. The patched orjson rejects deeply-nested inputs cleanly.Patches
Fixed in aiograpi 0.7.2 by migrating to
pyproject.toml(PEP 621) — single source of truth for dependencies. PyPI installs of 0.7.2 and later resolveorjson==3.11.8correctly.Workarounds
Force-install a non-vulnerable orjson alongside the affected aiograpi version:
Or just upgrade to a fixed aiograpi:
Resources
References