Vexa
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
Description from the maker (original English)
Let agents use temporary secrets without receiving their values in chat. Local stdio; Node 22+.
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Download the maker’s .mcpb file
The link goes to the maker’s own file; we do not host or repackage it. Open the file and Claude Desktop installs it.
Run this in your terminal
Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 7, 2026
The registry name was verified with the GitHub account geland, which owns the repo
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed –, and the latest version (0.0.2) reached the registry 17 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need MCPB on your computer.
In the same category: Vexa, emailmd, MCP WhatsApp Cloud
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
Render markdown into email-safe HTML, lint drafts for deliverability problems, and preview emails.
MCP server for WhatsApp Cloud API (Meta) — messages, media, and templates via Graph API
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
Give your agent access with a sealed secret.
Let your agent use temporary passwords, tokens, and one-time codes without putting their values in the chat. The local tool opens a private browser page, shows why the agent needs access, and applies secrets to requests you approve. No vault to configure. No Secret Handoff account.
This repository publishes the Claude Code and Codex plugins for the signed Secret Handoff 0.0.4 release. It includes the skill, local stdio MCP configuration, Claude detection hooks, and a status-only Claude card. Secret entry stays in your browser. These are repository plugins, not claims of vendor-curated directory approval.
Install CLI 0.0.4 first. Choose the archive for your platform and verify it with the release's SHA256SUMS. Run secrethandoff setup below to connect Claude Code and Codex when their CLIs are on PATH.
secrethandoff version
secrethandoff setup secrethandoff doctor
The expected version is 0.0.4. In the project where your agent works, run secrethandoff init to add credential guidance. Start a new agent session.
If you prefer to install the plugin yourself after installing the binary:
claude plugin marketplace add geland/secrethandoff-integrations
claude plugin install secrethandoff@secrethandoff
codex plugin marketplace add geland/secrethandoff-integrations
codex plugin add secrethandoff@secrethandoff
Restart the client or start a fresh session after installation. The binary must be on the PATH inherited by the client, including when launching a desktop app. secrethandoff doctor checks the local setup.
For a pinned marketplace, Codex accepts codex plugin marketplace add geland/secrethandoff-integrations --ref v0.0.4. Claude Code accepts the Git source https://github.com/geland/secrethandoff-integrations.git#v0.0.4 with claude plugin marketplace add. The release ZIP is also available for review.
Install the new binary first and verify that secrethandoff version reports 0.0.4. Both plugins launch the binary on your PATH.
claude plugin marketplace update secrethandoff
claude plugin update secrethandoff@secrethandoff
codex plugin marketplace upgrade secrethandoff
codex plugin add secrethandoff@secrethandoff
If your marketplace is pinned, switch its source to v0.0.4 first. Start a fresh agent session after updating. If you opted into Codex command approval, reapply the opt-in as described below.
Ask your agent: “Use Secret Handoff for the temporary credential this task needs.” Enter the value only on the Secret Handoff browser page. Approve a narrow API host, method, and path. The agent uses a reference and receives redacted results. Ask the agent to forget the secret when the task is done.
Secrets live in the tool process's memory until forgotten or the process exits. Filled secrets do not have an automatic expiry timer. Approved APIs receive the value. run_with_secret requires browser command approval by default; an approved command receives the value and can leak it. Owners can opt into verified Claude Code native permission prompts or Codex approval settings; Codex settings may authorize automatically. These modes do not change how secrets are entered.
Claude detection hooks can block some accidental pastes before they reach the model. They cannot erase a value already saved in the client's local transcript. Never paste secrets into chat. Read the security boundaries.
With the supported binary and stable Codex 0.155.1 in the 0.155 series, or the exact desktop build 0.162.0-alpha.17.2, the owner can choose native host-policy authorization. It follows the owner's settings, including allow rules, hooks, and automatic review; a human click is not required in that mode. Browser approval remains the distribution default. Secret entry always stays on the isolated Secret Handoff page.
From a checkout of this repository, reload the plugin, run the owner opt-in helper, then start a fresh chat. The helper is not included in the review ZIP:
codex plugin add secrethandoff@secrethandoff
python3 scripts/configure_codex_approval.py codex \ --plugin-dir ~/.codex/plugins/cache/secrethandoff/secrethandoff/0.0.4
The helper validates and edits only the installed mcp.json and .mcp.json launch environment. It does not change host policies, reviewer, or hooks. Reinstallation resets the opt-in; rerun it after reload. Use browser instead of codex to restore browser command approval. Agents may apply it only when explicitly instructed by the owner.
python3 scripts/check_codex_approval.py tests live allow/deny delivery with a harmless server. python3 scripts/check_codex_approval.py --installed-plugin checks that the real cached plugin starts with native mode and can list a new empty store. Both use child-only isolation and no secrets. Pass --codex-bin /absolute/path/to/codex to either check to validate the desktop bundled executable instead of the terminal CLI. Run python3 -m unittest discover -s tests -p test_codex_approval_config.py for helper preservation and refusal tests. Desktop prompt rendering needs a separate owner-run dummy test. Command results report client_policy / host_policy, never a claim that a human clicked.
request_secrets asks for 2–8 credentials on one local browser page. A group name identifies the request; fields have separate names and optional labels. One policy applies to every field, and all values become ready together. Wait using the group name. Forgetting the group or any member erases all members. Phone/headless relay fill is currently unsupported for groups. Existing request_secret calls still collect one value. This feature does not change command approval settings.
secrethandoff mcp.release-lock.json pins all six platform archives and six MCPB bundle hashes. Plugins start the binary on PATH; metadata does not replace or enforce that binary. Check secrethandoff version before use. Source and release provenance are available in the binary repository.Use Python 3.11+ and Node 24. Install requirements-review.txt in a virtual environment, then run:
python scripts/test_release.py
python scripts/validate.py --release python scripts/package.py --release python scripts/test_package.py
tests/card.test.tsx runs with the Claude mod test runner. tests/card-state.test.mjs runs with Node and TypeScript available through SECRETHANDOFF_TEST_TYPESCRIPT. See 0.0.2 release status, agent evals, and contributor guidance.