[dependabot]

Bumps aiohttp from 3.12.15 to 3.14.4.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)
  You can disable automated security fix PRs for this repo from the Security Alerts page.

---

Note

Medium Risk
Upgrades the async HTTP client used for API traffic; minor version jumps can affect connection, streaming, or TLS behavior even without SDK code changes.

Overview
Bumps the aiohttp dependency from 3.12.15 to 3.14.4 by updating pyproject.toml (^3.14.4) and regenerating poetry.lock (Poetry 2.5.1).

The lockfile refresh also pulls transitive updates, notably yarl 1.20.0 → 1.25.1 (required floor for aiohttp is now >=1.25.1). aiohttp and yarl now declare python-versions = ">=3.10", dropping Python 3.9 wheels—consistent with the project’s existing python = "^3.10" constraint.

No application or test source changes; behavior impact is limited to whatever changed in the upgraded HTTP stack at install/runtime.

^{Reviewed by Cursor Bugbot for commit ed36bc0. Bugbot is set up for automated code reviews on this repo. Configure here.}

[broly-code-security-scanner]

Broly Security Scan

Warning

Summary

13 actionable finding(s) in this PR

• 🔴 2 critical
• 🟠 5 high
• 🟡 6 medium

Main table: 12 shown, 1 omitted of 13 actionable finding(s) above the reporting threshold. See the repository Security tab for the full set.

Severity	Scanner	Issue	Location	Dismiss	Verdict
🔴 CRITICAL	SCA	black@25.9.0 — 3 vulnerabilities (worst: PYSEC-2026-2120)	poetry.lock:1	d1	🔺 TRUE_POSITIVE · Confidence: HIGH
🔴 CRITICAL	SCA	transformers@4.56.2 — 15 vulnerabilities (worst: PYSEC-2026-2290)	poetry.lock:1	d2	🔺 TRUE_POSITIVE · Confidence: HIGH
🟠 HIGH	SCA	urllib3@2.4.0 — 16 vulnerabilities (worst: GHSA-2xpw-w6gg-jr37)	poetry.lock:1	d6	🔺 TRUE_POSITIVE · Confidence: HIGH
🟠 HIGH	SCA	fsspec@2025.3.0 — 2 vulnerabilities (worst: GHSA-27vj-qcqg-25rc) → fixed in 2026.6.0	poetry.lock:1	d3	🔺 TRUE_POSITIVE · Confidence: HIGH
🟠 HIGH	SCA	pillow@11.3.0 — 36 vulnerabilities (worst: GHSA-vjc4-5qp5-m44j)	poetry.lock:1	d4	🔺 TRUE_POSITIVE · Confidence: HIGH
🟠 HIGH	SCA	virtualenv@20.34.0 — 10 vulnerabilities (worst: GHSA-p58f-9548-mpm2)	poetry.lock:1	d7	🔺 TRUE_POSITIVE · Confidence: HIGH
🟠 HIGH	SCA	pyarrow@21.0.0 — 2 vulnerabilities (worst: GHSA-rgxp-2hwp-jwgg) → fixed in 23.0.1	poetry.lock:1	d5	🔺 TRUE_POSITIVE · Confidence: HIGH
🟡 MEDIUM	SCA	datasets@4.1.1 — 2 vulnerabilities (worst: GHSA-379c-qx7v-6h59) → fixed in 5.0.1	poetry.lock:1	d8	🔺 TRUE_POSITIVE · Confidence: HIGH
🟡 MEDIUM	SCA	idna@3.10 — 2 vulnerabilities (worst: GHSA-65pc-fj4g-8rjx) → fixed in 3.15	poetry.lock:1	d10	🔺 TRUE_POSITIVE · Confidence: HIGH
🟡 MEDIUM	SCA	pytest@8.4.2 — 2 vulnerabilities (worst: GHSA-6w46-j5rx-g56g) → fixed in 9.0.3	poetry.lock:1	d12	🔺 TRUE_POSITIVE · Confidence: HIGH
🟡 MEDIUM	SCA	filelock@3.19.1 — 4 vulnerabilities (worst: GHSA-w853-jp5j-5j7f)	poetry.lock:1	d9	🔺 TRUE_POSITIVE · Confidence: HIGH
🟡 MEDIUM	SCA	requests@2.32.5 — 2 vulnerabilities (worst: PYSEC-2026-2275) → fixed in 2.33.0	poetry.lock:1	d13	🔺 TRUE_POSITIVE · Confidence: HIGH

---

Fix Suggestions

black@25.9.0 — 3 vulnerabilities (worst: PYSEC-2026-2120) — poetry.lock:1

Pin black to a patched version once available (e.g., poetry add black@latest --group dev) or, as a workaround, remove/upgrade black and run the formatter only in an isolated, untrusted-input-free CI environment; monitor PYSEC-2026-2120, PYSEC-2026-2121, and GHSA-3936-cmfr-pm3m for fixed releases.

transformers@4.56.2 — 15 vulnerabilities (worst: PYSEC-2026-2290) — poetry.lock:1

Upgrade to the latest transformers release that addresses the individual advisories (e.g., poetry add transformers@latest or poetry update transformers), and since no single fixed version covers all 15 advisories, review each advisory individually and pin to the newest available version while monitoring for patched releases; as an interim mitigation, avoid loading untrusted models/checkpoints and disable unsafe deserialization paths.

urllib3@2.4.0 — 16 vulnerabilities (worst: GHSA-2xpw-w6gg-jr37) — poetry.lock:1

Upgrade urllib3 to the latest available release (e.g., poetry add urllib3@latest or poetry update urllib3 to a version >= 2.5.0) and verify each of the 16 advisories individually against the release notes; if no fixed release yet covers all advisories, apply vendor-recommended workarounds (e.g., disabling affected features such as insecure redirects/certificate verification paths) and monitor for a patched release.

---

Dismiss false positives

Tick a box to dismiss the finding; untick it to bring the finding back. That is the same as replying /broly dismiss d1 and /broly undismiss d1. To record why it is a false positive, reply with /broly dismiss d1: your reason instead — Broly reuses those reasons to triage similar findings across the org.

• d1 · 🔴 CRITICAL · poetry.lock:1 · black@25.9.0 — 3 vulnerabilities (worst: PYSEC-2026-2120)
• d2 · 🔴 CRITICAL · poetry.lock:1 · transformers@4.56.2 — 15 vulnerabilities (worst: PYSEC-2026-2290)
• d3 · 🟠 HIGH     · poetry.lock:1 · fsspec@2025.3.0 — 2 vulnerabilities (worst: GHSA-27vj-qcqg-25rc)
• d4 · 🟠 HIGH     · poetry.lock:1 · pillow@11.3.0 — 36 vulnerabilities (worst: GHSA-vjc4-5qp5-m44j)
• d5 · 🟠 HIGH     · poetry.lock:1 · pyarrow@21.0.0 — 2 vulnerabilities (worst: GHSA-rgxp-2hwp-jwgg)
• d6 · 🟠 HIGH     · poetry.lock:1 · urllib3@2.4.0 — 16 vulnerabilities (worst: GHSA-2xpw-w6gg-jr37)
• d7 · 🟠 HIGH     · poetry.lock:1 · virtualenv@20.34.0 — 10 vulnerabilities (worst: GHSA-p58f-9548-mpm2)
• d8 · 🟡 MEDIUM   · poetry.lock:1 · datasets@4.1.1 — 2 vulnerabilities (worst: GHSA-379c-qx7v-6h59)
• d9 · 🟡 MEDIUM   · poetry.lock:1 · filelock@3.19.1 — 4 vulnerabilities (worst: GHSA-w853-jp5j-5j7f)
• d10 · 🟡 MEDIUM   · poetry.lock:1 · idna@3.10 — 2 vulnerabilities (worst: GHSA-65pc-fj4g-8rjx)
• d11 · 🟡 MEDIUM   · poetry.lock:1 · pygments@2.19.1 — 2 vulnerabilities (worst: PYSEC-2026-2987)
• d12 · 🟡 MEDIUM   · poetry.lock:1 · pytest@8.4.2 — 2 vulnerabilities (worst: GHSA-6w46-j5rx-g56g)
• d13 · 🟡 MEDIUM   · poetry.lock:1 · requests@2.32.5 — 2 vulnerabilities (worst: PYSEC-2026-2275)

Note

Re-scan this PR anytime with /broly scan — useful after /broly undismiss, or to refresh findings without a new push.

Broly — SAST (GLM-5.3-Flash) · Secrets · SCA · IaC · GH Actions · Base Images · Supply Chain Threats · Attack Hypotheses · Adversarial Verification

We're continuously improving Broly's accuracy and finding quality — your feedback is valuable. False positives, missed findings, bugs, and feature requests all welcome.