Read 2026-10-23
crewaiinc/crewai: technical due diligence, read from its repository
What Borehole read in crewaiinc/crewai's code and history at 07361f6: its licence, security advisories, releases, review practice, tests, dependencies and operations. Every finding cites the file and line it rests on. The full report also covers the team and asks the questions the code cannot answer.
| Dimension | Rating |
|---|---|
| Architecture & Codebase — including AI provenance | strong |
| Engineering Process & SDLC Maturity | adequate |
| Product & Engineering Maturity | adequate |
| Security & Reliability Posture | adequate |
| Observability & Operations | adequate |
| Fit with the Acquirer | not assessable |
What the repository says
.github/CONTRIBUTING.md:3: states a policy on AI-generated contributions
Architecture & Codebase — including AI provenance
stronglib/crewai/src/crewai/flow/runtime/__init__.py is 4,071 lines, 38× this codebase's median file
adequateThe median source file here is 105 lines. 32 files exceed 1,050 lines, the largest being 4,071. Size is not a defect on its own — some problems genuinely live in one place — but these are where merge conflicts, review fatigue and single-owner knowledge concentrate, and they are the first thing that slows an inheriting engineer down.
lib/crewai/src/crewai/flow/runtime/__init__.py:1 — 4,071 lines (38× median) |
lib/crewai/src/crewai/telemetry/tracing/handlers.py:1 — 3,773 lines (35× median) |
lib/crewai/src/crewai/agents/agent_executor.py:1 — 3,549 lines (33× median) |
lib/cli/src/crewai_cli/crew_run_tui.py:1 — 3,422 lines (32× median) |
lib/crewai/src/crewai/llms/providers/openai/completion.py:1 — 3,126 lines (29× median) |
Engineering Process & SDLC Maturity
adequate10% of recent commits change more than 2,000 lines
adequate170 of 1773 commits in the last 24 months changed over 2,000 lines of code, not counting reformatting, renames, vendored or generated code. A change that size is rarely reviewed line by line. Ask how they were reviewed, and what made them that large.
| 170 commits over 2,000 lines |
Product & Engineering Maturity
adequate17,074 unresolved debt markers, 25.5 per thousand lines
adequateThe codebase carries 17,074 TODO, FIXME, HACK or XXX markers across 669,292 lines. These are the team's own notes about work they knew was unfinished. A high density is not automatically bad — it can mean an honest team — but every one of them is a decision deferred, and after an acquisition they are deferred onto someone who was not in the room.
.github/workflows/vulnerability-scan.yml:107 — # TODO: drop this ignore when bumping nltk past 3.10.3 to a patched |
.github/workflows/vulnerability-scan.yml:116 — # TODO: drop this ignore when bumping accelerate past 1.14.0 to a |
.pre-commit-config.yaml:32 — # TODO: drop --ignore-vuln [blanked] when bumping nltk past 3.10.3. |
.pre-commit-config.yaml:33 — # TODO: drop --ignore-vuln [blanked] when bumping accelerate past 1.14.0. |
conftest.py:288 — "authorization": "[blanked]", |
Security & Reliability Posture
adequate1 deserialiser that can run code
adequate1 places read data with a deserialiser that executes what it reads (pickle). That is safe on data the system wrote itself and remote code execution on anything else. Ask where each one's input comes from.
lib/crewai/src/crewai/utilities/file_handler.py:167 — code-executing deserialiser |
7 locked dependencies have published vulnerabilities, 34 high or critical in what ships
adequate7 of 498 locked dependencies match 51 published advisories in OSV, 37 of them rated high or critical. A lockfile says which version is installed, not whether the flawed code is reachable from this product, so each needs a look. Most are fixed by an upgrade, and the advisory names the version. This lockfile pins development tools alongside what ships and does not say which is which, so some of these may never reach a user.
5 advisories against this project in the last 24 months, 4 high or critical
adequateThe newest was published on 2026-10-19. A project that publishes advisories is telling its users about its fixes, and the count and pace say how much security work it carries. Whether 2 of them are fixed in the code you ship could not be told from this history: its version is only known to be at least 1.15.29. Answer lives in: the release you run, set against each fixed version. The record is incomplete: NVD: 5 records name crewAI with nothing tying them to this repository (CVE-2026-2275, CVE-2026-2285, CVE-2026-2286, …).
| 2026-10-19: CVE-2026-112308 (high): crewai-tools through 1.15.28 contains a server-side request forgery vulnerability in SeleniumScrapingTool that allows attackers to fetch internal resources beca |
| 2026-09-13: CVE-2026-37008 (high): CrewAI before fb2323b offers a Python blocklist approach that operates at the wrong level of abstraction, a different vulnerability than CVE-2026-2275. Import-t |
| 2026-08-27: CVE-2026-37007 (critical): A vulnerability in FileWriterTool in crewai-tools <= 1.10.2rc1 allows a remote attacker to achieve code execution via malicious path traversal sequences in the |
| 2026-08-27: CVE-2026-37009 (moderate): A SQL injection vulnerability in NL2SQLTool in crewai-tools v1.10.2rc1 allows a remote attacker to execute arbitrary SQL commands via an unsanitized sql_query a |
| 2026-07-13: GHSA-mr4r-hcgx-8p4h (high), fixed in 1.15.1: crewai-tools SSRF redirect bypass exposes internal services |
Observability & Operations
adequateSends usage data to telemetry.crewai.com by default
adequatelib/crewai-core/src/crewai_core/telemetry.py sends to telemetry.crewai.com, and lib/crewai-core/src/crewai_core/telemetry.py leaves reporting on as shipped. Every installation reports to the vendor unless its operator turns this off. For a buyer running it inside their own product, that is an outbound data flow to a third party: ask what is sent, how often, and whether the buyer's customers were told.
lib/crewai-core/src/crewai_core/telemetry.py:45 — reports to telemetry.crewai.com |
lib/crewai-core/src/crewai_core/telemetry.py:163 — the reporting setting is on by default |
Fit with the Acquirer
not assessableNot rated: structurally not assessable from a repository.
Read more, or read your own
The full report on crewaiinc/crewai: every dimension, the questions for the room, and the audit log.
Survey a repository: free on public repositories. Private code is read on your own machine by a collector whose source you can read first.