Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
There was a problem hiding this comment.
All reported issues were addressed across 2 files
Reply with feedback, questions, or to request a fix.
Fix all with cubic | Re-trigger cubic
Greptile SummaryThe PR consolidates SOC 2 and other certification data into one Compliance fact across all comparison profiles.
Confidence Score: 5/5The PR appears safe to merge. No blocking failure remains.
|
| Filename | Overview |
|---|---|
| apps/sim/app/(landing)/comparisons/comparison-sections.ts | Replaces the two former certification rows with one type-checked Compliance row consumed by the table and JSON-LD builder. |
| apps/sim/app/(landing)/comparisons/utils.ts | Updates the security FAQ to read the consolidated compliance facts for Sim and each competitor. |
| apps/sim/lib/compare/data/types.ts | Replaces the obsolete security fact fields with the canonical compliance field. |
| apps/sim/lib/compare/data/sim.ts | Lists Sim’s SOC 2 Type II, ISO 27001:2022, and GDPR certifications under one Trust Center-backed fact. |
| apps/sim/lib/compare/data/competitors/stackai.ts | Preserves StackAI’s qualified SOC 2 status in both the full and visible consolidated compliance values. |
| apps/sim/lib/compare/data/competitors/tines.ts | Preserves Tines’s SOC 2 status in both the full and visible consolidated compliance values. |
| apps/sim/lib/compare/data/competitors/vellum.ts | Preserves Vellum’s SOC 2 status in both the full and visible consolidated compliance values. |
| apps/sim/lib/compare/data/competitors/power-automate.ts | Preserves Power Automate’s environment-qualified SOC 2 scope in both consolidated values. |
| apps/sim/lib/compare/data/competitors/microsoft-copilot.ts | Preserves Microsoft Copilot’s SOC 2 status and scope in both consolidated values. |
| apps/sim/lib/compare/data/competitors/crewai.ts | Preserves CrewAI AMP’s SOC 2 Type 1 status and scope in both consolidated values. |
| apps/sim/lib/compare/data/competitors/claude-cowork.ts | Preserves Anthropic’s company-wide SOC 2 status and Cowork-specific qualification in both consolidated values. |
Reviews (4): Last reviewed commit: "improvement(comparisons): remove obsolet..." | Re-trigger Greptile
c3825bf to
f611370
Compare
|
@cubic-dev-ai review this PR |
@j15z I have started the AI code review. It will take a few minutes to complete. |
f611370 to
661020f
Compare
|
@cubic-dev-ai review this PR |
@j15z I have started the AI code review. It will take a few minutes to complete. |
661020f to
3ca2ce8
Compare
|
@cubic-dev-ai review this PR |
@j15z I have started the AI code review. It will take a few minutes to complete. |
1a298a7 to
3aa79cf
Compare
3aa79cf to
8926217
Compare
A profile stated its certifications twice: once as a bare SOC 2 yes/no and again under additional compliance. Merge them into one `compliance` fact that carries every certification, its report scope, and its access conditions in one place, keeping the long-form text from both. Add a `scim` row next to SSO. Sim ships SCIM 2.0 directory provisioning on Enterprise, and provisioning is a distinct buying question from authentication, so it deserves its own comparable row rather than a clause buried in the SSO cell. Where a profile described provisioning inside its SSO fact, that content moves to the new row and SSO keeps authentication. Every SCIM value is sourced to a vendor page fetched today. Where a product does not publish its provisioning support, the row says it is not documented and carries `confidence: 'unknown'` rather than asserting the product lacks it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Runs the comparison-page fact check end to end with no arguments: preflight that web fetching and the browser both work, one read-only checker per profile, then a fresh fixer that sees only that profile's flags and decides for itself which are real. Encodes what three review rounds taught us. Long-form value and detail are never shortened, because answer engines read them and a previous run cut 456 KB of sourced text. Absence of evidence is written as "not documented", never as absence of the feature. Pages that defeat fetching are retried in a browser and then reported as unverifiable rather than guessed. Dates refresh only for pages actually read. Nothing is written to the repo except edits to the data files. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Recheck all 20 competitor profiles and Sim's own against public primary sources in batches of 5 (checker + independent fixer per profile), correcting stale capability, pricing, licensing, and product lifecycle claims while preserving explicit uncertainty and scope qualifications. Notable corrections: Flowise wound down and archived its repo; LangChain now holds ISO 27001 certification the profile said it lacked; OpenAI AgentKit's ZDR/background-mode compatibility was stated backwards; Gumloop repositioned from visual builder to chat-agent product and dropped its free tier; Retool self-hosting is Enterprise-only, not all tiers; Tines audit-log retention is now configurable, not fixed; Power Automate is SOC 2 GCC High in-scope, contrary to the prior claim. Refreshed Sim's own internal counts cited across every profile: blocks 266->356, tool actions ~3,900->~5,500, providers 21->25, KB connectors 51->65, trigger apps 61->64. Claims that could not be independently confirmed are marked confidence: 'unknown' rather than asserted, and the sources arrays that support each claim were refreshed to the date they were actually re-read: docs.sim.ai, trust.sim.ai, and each competitor's own docs, security/trust center, and pricing pages, per the audit-comparisons skill. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
8926217 to
17545d5
Compare
Summary
soc2andadditionalComplianceinto a singlecompliancefact so a profile states its certifications once, with a migration test asserting no detail was dropped in the move.Scope
This branch is the data layer only, plus the three call sites that had to change to accept the new shape: the security rows in
comparison-sections.ts, the compliance key in the security FAQ, andprose-text.tsxrendering a segment's source URL.The comparison page redesign and the UI that renders these citations are a separate change on
feat/redesign-comparison-page. Both branches are based on staging independently. Whichever lands second will need a small conflict resolution inutils.tsand[provider]/page.test.tsx.Review
The facts went through two rounds of AI-assisted citation review (checker + independent fixer, per profile, batches of 5), and the corrections from those rounds are applied here. Claims that could not be confirmed are labeled as unknown or estimated rather than asserted — see Unverifiable claims below for the specific list. This is not a claim that every claim and link was independently confirmed — access limitations and unresolved items are recorded there.
Notable corrections from this pass: Flowise wound down operations and archived its GitHub repo (Aug 2026); LangChain achieved ISO 27001 certification that the profile said it lacked; OpenAI AgentKit's Zero-Data-Retention/background-mode compatibility was stated backwards; Gumloop repositioned entirely from visual workflow builder to chat-based agent product and dropped its free tier; Retool self-hosting is Enterprise-plan only, not available on all tiers as previously stated; Tines audit-log retention is now org-configurable (30 days–2 years), not fixed at 2 years; Power Automate is SOC 2 Type 2 in-scope for GCC High, contrary to the prior claim. Sim's own internal counts, cited across every competitor profile, were stale and refreshed: blocks 266→356, tool actions
3,900→5,500, LLM providers 21→25, knowledge-base connectors 51→65, trigger app coverage 61→64.Unverifiable claims (confidence:
'unknown', or otherwise flagged as access-limited)These rest on a source that could not be read (paywall, dead link, bot-blocked, or no longer contains the claim) after both a direct fetch and a browser retry. They are marked accordingly in the data rather than asserted as fact — do not treat them as confirmed:
anthropic.skilljar.comis now a bare login gate with no visible course content.platform.versionControlDepth,observability.tracingDepth,observability.durabilityModel,observability.failureAlerting: all four cited doc pages now redirect to a generic overview page with none of the original content, and no replacement page was found.support.sla's secondary citation (windowsnews.ai→windowsforum.com) still returns 403./blogsection 302-redirects to/docs), and no independent corroboration exists.support.companyMaturity's Crunchbase citation (crunchbase.com/organization/workato) returns 403 / a Cloudflare challenge; the fact is still backed by two other sources (TechCrunch, Tracxn) that were independently verified, so it isn't wholly unsupported, but that one citation itself couldn't be read.Two additional flagged-but-resolved items worth noting, since they were close calls:
platform.subWorkflowsandplatform.customBlocks, which may repeat a stale "Lobster only has run/pipeline/approval steps" claim discovered as a side effect of fixing a different flag in the same file.observability.executionLimitsandsupport.supportChannelsmay still describe free-tier concurrency/support figures for a Free plan that Gumloop has since removed from its pricing page; this was noted but out of scope for the flags actually raised.Type of Change
Testing
bunx vitest run "app/(landing)/comparisons" "lib/compare")tsc --noEmitclean across the appChecklist
🤖 Generated with Claude Code