v0.0.14-beta.2: Safer Policies, Config Improvements, and CLI Reliability

Policy safety improvements

• Strengthened block-rm-rf to detect destructive deletion targets including $HOME,
  ~/..., critical absolute paths, find -delete, and unresolved shell expansions.
• Improved force-push detection to catch bundled flags such as -fu, while allowing safer
  variants like --force-with-lease.
• Added an option to disable custom policies without renaming or deleting policy files.

Setup and CLI improvements

• Fixed the config wizard crashing when “Just this project” included assistants without
  project-level configuration support.
• Added clearer custom-policy status to the policy selector.
• Fixed the dashboard “install all” command to generate valid CLI syntax.
• Refined terminal branding, colours, and NO_COLOR support.
• Renamed and improved the CLI integration suite for more reliable local and CI testing.

Fixes

• Fixed duration formatting around minute and hour boundaries.
• Fixed dashboard copy-button feedback during rapid clicks and component unmounts.
• Removed checkout-directory assumptions from Pi integration tests.
• Improved translated documentation validation and automatic retry handling.
• Clarified AgentEye SDK payload serialization documentation.

v0.0.14-beta.0: New CLIs, Copilot Fix, and Failproofai Config

New CLI integrations (now 12 total)

Five new agent CLIs added, each with full real-time enforcement plus offline audit/dashboard support:

• Goose (Block): uses its Open Plugins hooks spec, deny only on PreToolUse, session parsing from its SQLite DB
• Antigravity (agy): its own named-hook contract, not a Claude clone, camelCase protojson payloads, unique response shapes per event
• Devin (Cognition): a pure Claude-clone contract, verified deny behavior across every event including Stop
• Factory (droid): Claude-compatible but with two schema quirks: no "hooks" wrapper, and deny via exit code 2 instead of JSON decision (except on Stop)
• OpenClaw: uses in-process plugin hooks since its file-based hooks are observation-only; ships a static plugin package, user-scope only

Gemini CLI integration removed entirely, since Google retired it in favor of Antigravity. All Gemini-specific code, maps, and dashboard providers were stripped out.

Setup experience overhaul

• New failproofai config command: a guided 4-step wizard (Where, Assistants, Policies, Review) replacing manual flag setup, fully rebranded with the befailproof.ai pixel logomark and pink/teal palette
• First run now flows setup, audit, dashboard automatically in one command
• Wizard's guard-selection step is now multi-select (bundles combine instead of single-choice)
• New "Everything available" shortcut for assistants, mirroring the policies option
• Dashboard launch splash rebranded to match the wizard

Key fixes

• Copilot CLI: file policies were silently letting file access through due to drifted hook input keys; now fixed and verified with a live .env read denial
• Antigravity: file writes were slipping past path/content policies due to wrong tool names; corrected and verified
• Devin: session transcripts were rendering duplicated 2 to 4x due to its forest-structured message storage; now reconstructs the real conversation path
• Dashboard header made responsive (was breaking at narrower widths)
• Stop-event allow-notes no longer leak into agent-facing stdout, fixing a message wall on droid

Docs

• Docs site restructured with AgentEye and FailproofAI as top-level tabs
• README's CLI logo grid made responsive (2x6 table)
• Documented that VS Code Copilot agent mode is already covered by existing copilot/claude integrations

v0.0.13 : Hermes integration and hook reliability

What’s new

• Added Hermes (by @NousResearch) integration for real-time policy enforcement, audit replay, project history, session viewing, and JSONL downloads.
• Added Hermes support across the README and documentation.
• Synced supported hook events with the latest Codex, Copilot, and Claude specifications.
• Added the automated hook-sync container for detecting upstream CLI hook changes.

Fixes

• Skip workflow completion gates during Claude Code plan mode.
• Correct hook timeout units for Claude, Codex, and Cursor.
• Repair Codex configurations containing the obsolete top-level version field.
• Support Pi ≥0.80 in live install/uninstall roundtrip tests.
• Fix translated documentation regressions.

Dependencies

• Added yaml for comment-preserving Hermes configuration updates.
• Added sql.js as the SQLite fallback on older Node.js versions.
• Included the dependency and GitHub Action updates shipped throughout the beta cycle.

0.0.13-beta.2: Fixes

Fixes

• Skip the require-*-before-stop workflow gates during Claude Code plan mode (permission_mode: "plan") — plan mode makes no commits/pushes/PRs by design, so the gates were wrongly demanding actions plan mode forbids (e.g. git push with nothing to push, blocking the agent from finishing). (#488)
• Fix translation regressions from the #486 docs sync: strip hallucinated {#…} heading-ID syntax that broke MDX parsing (ja/built-in-policies), restore dropped YAML frontmatter (tr/cli/hook) and the --cli …|hermes inline code span (he/configuration), and un-translate an inline-code placeholder (de/cli/auth). (#487)

v0.0.13-beta.0: Synced Hooks Schema

Syncs hook integrations with the latest CLI docs — adds newly-documented events for Codex, Copilot, and Claude. Fixes a Claude & Cursor timeout bug (was 60000 in a seconds field ≈ 16.7h; now 60 = 60s).

v0.0.12 - Codex hooks: drop invalid version field + correct timeout unit (seconds)

v0.0.12 - Codex hooks: drop invalid version field + correct timeout unit (seconds)

v0.0.12-beta.0 — Codex hooks: drop invalid version field + correct timeout unit (seconds)

docs(changelog): backfill 0.0.12-beta.0 with dependency bumps + trans…

v0.0.11 — Introducing Audit

First stable release of the 0.0.11 line. Published to npm as failproofai@0.0.11 under the latest tag.

Highlights since 0.0.10

Features

• /audit dashboard + failproofai audit CLI — scan your agent-CLI session history and get a personality-driven audit (8 archetypes, 0–100 score, shareable poster) with prescribed policies. Runs fully local; install-free via npx -y failproofai audit.
• Email-OTP auth — failproofai auth login | logout | whoami, wired to the api-server; tokens stored at ~/.failproofai/auth.json.
• Supply-chain CI gate (OSV-Scanner + Socket) and a PR-level MDX parse check.
• Broader, more reliable PostHog telemetry across all four channels.

Breaking changes

• Default policy namespace renamed exospherehost → failproofai. Configs that reference builtins as exospherehost/<name> must update to failproofai/<name>. Flat shorthand (e.g. "sanitize-jwt") is unaffected.
• Removed the legacy cloud auth + event-relay subsystem (the old failproofai login / logout / whoami / relay / sync commands) ahead of the new auth surface above.

Other

• Community moved from Slack to Discord; docs live at https://docs.befailproof.ai/introduction.

Full per-change detail is in the CHANGELOG.

v0.0.11-beta.12 — Discord community link, responsive activity table + mode-pill fix, audit copy

Features

• Collapse the dashboard "Reach Us" dropdown's three GitHub links (Request a Feature / Report an Issue / Ask a Question) into a single Feedback & Issues entry pointing at the GitHub issue chooser (/issues/new/choose).
• Reorder the policies → activity table columns to: time · decision · event · cli · tool · policy · reason · mode · duration · session.

Fixes

• Fix the policies → activity table collapsing on narrow / portrait windows. Columns no longer overlap — each data cell clips with an ellipsis at its own edge and headers stay on one line — and the table holds a readable min-width (1280px), scrolling horizontally below that via a themed scrollbar instead of squeezing columns into each other. The badge / long-header columns (decision, event, cli, mode, duration, session) were widened so their content fits — the mode column in particular now holds its widest pill (bypassPermissions) instead of clipping it mid-word, and the mode pill truncates with an ellipsis + hover tooltip if a longer / custom mode ever appears.

Docs

• Replace the community Slack invite with Discord (https://discord.gg/2zjBZP7yQJ) everywhere it's user-facing: the failproofai --help LINKS banner, the dashboard "Reach Us" dropdown, and the README community badge (English + 14 translations). The Slack webhook notification example (examples/policies-notification.js) is intentionally left as-is — it's a feature integration, not a community link.
• Reword the /audit invite card ("Share with friends" / "wanna know how your friends' agents score?") and grammar-pass the X/LinkedIn share templates (article/adverb/coordination/comma-splice fixes only — no behavioral or structural change).

v0.0.11-beta.11 — invite emails show audit score, quieter server logs, PR-level MDX parse check

Features

• Add a PR-level MDX parse check (bun run validate:mdx, wired into the CI docs job) that compiles every docs/**/*.mdx with the same MDX engine Mintlify runs at deploy time. mintlify validate only checks docs.json structure and nav links — it never parses page content — so syntax errors slipped through to the post-merge deploy. This catches them on the PR instead (#455).
• Invite emails now include the inviter's audit score: sendInvites() and the /api/audit/invite proxy forward a clamped (0–100) score to the api-server, which renders "my agent scored a N/100" in the body. Threaded AuditDashboard → ComeBackBetterSection → InviteDialog; optional end-to-end so it degrades to score-free copy when absent (#456).
• Rewrite the X/LinkedIn share templates (10 each): lead on the score and archetype and end on the npx -y failproofai audit CTA + handle (@failproofai / @Failproof AI), with no URLs in the copy so the pasted audit-card image isn't replaced by a link-preview card (#456).
• Enlarge the audit poster's four corner labels so they read clearly at share size, on both the dashboard render and the downloaded PNG (#456).

Fixes

• Fix three translated docs pages that failed the Mintlify deploy parse. docs/tr/cli/audit.mdx had a dropped closing backtick that pushed <slug> out of its inline-code span (parsed as an unclosed JSX tag); docs/ja/built-in-policies.mdx and docs/zh/built-in-policies.mdx carried translator-injected {#id} heading anchors that MDX reads as JS expressions. All three now match the other 12 locales (#455).
• Stop the failproofai server log from repeating the benign Next.js "Failed to find Server Action" deployment-skew error. A browser tab left open across a dashboard rebuild/upgrade POSTs a stale Server Action ID; the client recovers via Next's graceful 404, but the standalone server still logged a 3-line error block to stderr per stale request. The start launcher now pipes the server's output through a filter (scripts/skew-log-filter.ts) that drops just that block — all other output, and color via FORCE_COLOR, passes through untouched; dev is unchanged (#456).

Dependencies

• Add @mdx-js/mdx as a dev dependency, used by the new validate:mdx docs parse check (#455).

---

Published to npm under the beta dist-tag — install with npm i failproofai@beta or run npx -y failproofai@beta.