← Blog
SaaS integrations
Give AI the context to continue the work: source conversations, open tasks, and unresolved details, from adoption decisions to integration code.
SaaS integration guidesCircleback webhooksDub UTMSaaS workflow automation
23 articles
Claude Code 2.1.325: Sandboxed Bash Prompts and Skill Approval
Review sandbox autoAllow, scoped ask and deny rules, and scheduled Skill approval. Read-only audit with 729 synthetic cases; CLI comparison unverified.21 minClaude Code Permission Wildcards: What * Allows and How to Audit
Compare Claude Code 2.1.322 and 2.1.323 in isolated tests: broad Bash wildcards, spaces, variable warnings, and a read-only settings audit.25 minClaude Code Subagent MCP: Folder Trust and Startup Controls
Cloned repo inline MCP waits for folder trust. Tests on Claude Code 2.1.319 cover 75 isolated cases, strict and bare exceptions, and a read-only audit.22 minPostman Auto-Install: Check Claude Code and Codex Settings
Postman auto-install reports, rollback response, source changes and a read-only settings checklist. Local checks; removal and exfiltration unverified.23 minClaude Code Opus 4.8 Fast Mode Sunset: Settings and Migration
Opus 4.8 fast mode is scheduled for removal on 2026-11-30 with no standard-speed fallback. Official guidance, CLI 2.1.317 checks and 48 read-only cases; migration and billing untested.19 minNetlify Free Plan Credits: Deploy Limits and Paused Sites
Netlify credit-based Free has 300 monthly credits and 15 per production deploy. Twenty deploy charges assume no other usage. Rates checked Nov 5, 2026; pause options, old limits and Pages compared. Live usage and recovery untested.15 minCircleback MCP with Claude Code: setup, read-only options, and data access
Connect Circleback MCP to Claude Code, inspect OAuth and project approval, and retrieve meeting context. A 401 experiment separates a user-only scope candidate from verified read-only access, history limits, and output limits.19 minClaude Code context handoff: compact, clear, resume, and revision checks
Official excerpts and 21 Claude Code 2.1.308 trials check handoff loading, resume, fork, and another folder. Includes JSON validation, SHA-256 limits, compaction guidance, and untested conditions.35 minWhich CLAUDE.md and AGENTS.md files load? A measured monorepo
Measure Claude Code 2.1.308 and Codex 0.163.0 with a synthetic monorepo: nested loading, ancestors above Git root, four import hops, and conditional AGENTS.md support, with events and canaries.19 minClaude Code MCP configuration: .mcp.json and local, project, user scopes
Official excerpts and real Claude Code 2.1.308 output explain MCP scope precedence, pending approval, folder differences, missing type, header order, and duplicate removal on macOS.28 minClaude Code Hooks: Exit 1 vs 2 and Hooks That Do Not Block
20 cases measured with Claude Code v2.1.308 and Haiku on macOS: PreToolUse, PostToolUse, onFailure, JSON decisions, unmeasured conditions, and an offline exit-status demo.13 minClaude Code /usage: Breakdown, Session Cost and Zero Results
Read Claude Code /usage without confusing local session cost with plan limits. Six isolated 2.1.308 probes compare usage, cost and stats, with output and caveats.12 mincodex exec Skips MCP? 0.169 Untrusted Projects and --trust
Tested Codex CLI 0.169.0 in 21 isolated, unauthenticated cases: project MCP, config trust, persistent --trust, explicit untrusted, and CI decisions.16 mincodex exec: stdin, JSONL, Sandboxes and Exit Codes Tested
Seven Codex CLI 0.163.0 probes cover stdin, stdout/stderr, JSONL, schema output, read-only, workspace-write and resume, with a documentation-based claude -p comparison.21 minAGENTS.md Security in Downloads: Claude Code and Codex Tested
Test Claude Code 2.1.312 and Codex 0.163.0 outside Git with harmless canaries, noninteractive trust limits and a read-only configuration inventory.15 minOne SKILL.md for Claude Code and Codex: Paths and Startup Cost
Tested Claude Code 2.1.312 and codex-cli 0.163.0 in a temporary HOME: SKILL.md paths, startup description cost, and symlinks. Automatic selection and real turns are untested.19 minBlocking Claude Code Writes Outside the Working Directory: Bash and Codex Tested
permissions.blockWritesOutsideWorkingDirectories measured on Claude Code 2.1.316: Write/Edit refused, Bash not stopped. Compares sandbox settings and Codex 0.166.0. Linux/Windows, MCP, hooks and more remain unverified.34 minClaude Code and Codex in Parallel: git worktree and .worktreeinclude
Worktree creation and .worktreeinclude tested with Claude Code 2.1.312 before login, plus Git edit isolation and documented codex --cd. Parallel model execution untested.15 minDub conversion tracking: lead and sale APIs with Stripe customer mapping
Connect Dub clicks to signups and sales. Check Business requirements, lead and sale fields, USD/JPY amounts, invoiceId, and Stripe metadata with official sources and a credential-free JSON planner.14 minDub webhook signature verification: raw bodies and replays in Node.js
Verify Dub-Signature over the received body. A Node.js/Bun Request experiment tests 36 body/signature pairs, second-read TypeErrors, and a UTF-8 BOM limit before a minimal AI handoff.16 minRecord Slack huddles and feed AI at 22:00: six Circleback settings
A Circleback workflow for giving AI usable work context: Slack huddles, a 22:00 sync, source statements, open tasks, and adoption criteria. Separates workflow experience, documented features, and local test evidence.37 minCircleback action items to tasks: testing webhook signatures, owners, and replay keys
Runnable Circleback webhook code tests signatures, owners, SQLite upserts, and stale replays. A separate status example preserves destination-owned DONE. Uses synthetic data.33 minTrack Dub referral links in Japanese and English without breaking referral IDs
Adding UTMs can re-encode existing referral parameters even when their decoded values stay the same. Reproduce spaces, tildes, and duplicate keys locally, with limits for signed URLs and commission attribution.21 min