Agentalog bronze

search_bundles

Search the index of OKF bundles: free text over name, tagline, description and origin; filter by provenance (github|domain), repository, version, language, license, root type or concept count; search root concept text. Paginate with limit/offset. Only `live` bundles. Free text matches the name, the tagline, the description and the origin identifier (`owner/repo:path` or the bundle URL). No `total` on purpose: `GET /api/okf/stats` has it.

get_bundle

One bundle's card: root index.md URL, version, concept count, provenance and repository signal. `live` and `low` (example or fixture bundles kept out of the search) both answer here.

index_stats

Size of the index by provenance and when it last changed.

listing_dependencies

Read a tool's declared package/repository/domain chain and cached public source evidence. No purchase or collection. Repeated registry, repository, registration and DNS lookups, comparison code and an alert process that must keep running after the conversation. A declaration does not prove ownership. Source failures and missing dates remain unknown. Collected PyPI provenance is not independently verified. No measured token or money saving is claimed.

watch_dependency_changes

Use an existing paid watch pass to follow package, publisher, repository and domain changes. Never purchases a pass. The listing must already be in your watchlist. No separate purchase. Sources refresh every six hours through shared caches. Source quotas can defer a check. History records before, after, source and observation date. Domain expiry alerts use 30, 15 and 7 days when the registry publishes a date. Missing dates and source failures remain unverified. Account holders receive operational e-mail; wallet owners read API history. No account-recovery or takeover attempt is made.

dependency_changes

Read your watched tool's before/after evidence, dates and sources. Continue through history_next.

feed_find

Does this site publish a feed? Answers with every feed known for a domain — the question a reader or an agent actually asks. Find feeds by text, format, language and availability. Filter and text search run in the SAME full-text query: values inside one parameter are OR'd, different parameters are AND'd. Without `q` the order is by newest in the index (`recentes`) and `next_cursor` is a keyset — follow `next` to walk everything. With `q` the order is relevance and paging is by `offset`. We index and link; the feeds belong to whoever publishes them.

feed_posts

What has this feed published? The posts of one feed as our queue read them — title, date, author, summary, categories and enclosure, without the body. Find the feed first (feed_find or feed_search); this takes an id from this index, never a URL. Up to 40 posts, in the order the feed publishes them, WITHOUT the body — the body of one post comes from `/items/:ord`. It is OUR copy, read on our own schedule — not the live feed and not a proxy: the only thing this endpoint accepts is an id from this index, never a URL. `lido_em` says when we read it. Empty list when the feed has not been read yet.

feed_post

Read one post whole: the body in Markdown (`conteudo`) and the author's formatting as safe HTML (`conteudo_html`), plus the link to the original. `ord` comes from feed_posts; 0 is the newest. Two bodies for the same post, both from the copy we stored when we READ the feed — a request never fetches the site: `conteudo_html` keeps the author's formatting through an allowlist (only known tags and attributes survive; script, style, iframe and event handlers are dropped with their content), and `conteudo` is the same text as Markdown — the shape an agent usually wants. Neither can carry script. 404 when the feed or the position is unknown.

feed_search

Search the feed index by text and filters: format, audio/video, community, namespace and year of the newest item. Find feeds by text, format, language and availability. Filter and text search run in the SAME full-text query: values inside one parameter are OR'd, different parameters are AND'd. Without `q` the order is by newest in the index (`recentes`) and `next_cursor` is a keyset — follow `next` to walk everything. With `q` the order is relevance and paging is by `offset`. We index and link; the feeds belong to whoever publishes them.

mcp_index_signals

Does this MCP server still answer, with which access, and is its repository alive? What independent MCP indexes measured about one listing of this catalog, each credited with its license. Take the id from list_listings. A signal belongs to a listing only when it is the SAME object: the name in the official MCP registry, the exact remote endpoint URL or the GitHub `owner/repo`, in that order. A shared domain never matches — two servers on one host are different products. Up to 20 signals, one per index and record.

mcp_index_search

Browse what independent MCP indexes measured, including servers this catalog does not list yet (casou=0). Newest change first. `casou=0` is what the indexes know and the catalog does not — the list to look at before submitting a server. Filters are AND'd. With `q`, `total` is the exact number of matches, including when the requested page is empty.

check_tool

Before installing a remote MCP server, check its address: signals from the address itself (consumer tunnel, no TLS), the catalog listings with that address or host and their security level, and facts about its domain. Never visits the address. Send the endpoint URL. The answer never contacts that address: it reads what our weekly visits measured and what our domain data knows. `sinais_do_endereco` comes from the address alone (consumer tunnel, `http://` without TLS). `registros` lists the catalog listings with this exact address first (`exato: true`), then up to 20 others on the same host, each with its level and signals; `mesmo_host` counts them all. `dominio` has the facts about the host's domain. An address outside the catalog answers `no_catalogo: false` and `registrar`: submit it to get it reviewed. Cached for 5 minutes.

listing_security

Read a remote MCP server or SKILL.md review before installing. MCP: server, domain, tools and measured signals. Skill: static findings with file, line, excerpt and context; no command execution or referenced-script review. Unread skills stay unchecked. Read the limits in aviso. Take the id from list_listings. We visit remote MCP servers: `initialize` and `tools/list`, without credentials or tool calls, again 7 days after a success. A failure needs a second check. MCP signals: dns_morto, sem_https, tls_invalido, tunel, nao_responde, nao_mcp, texto_suspeito, redireciona_fora, repo_arquivado, repo_parado, dominio_recente, login, bloqueado. For skills, `tipo: skill` identifies a static SKILL.md review. It flags secret reads, download-and-execute commands, external uploads, invisible characters, long encoded text and broad allowed-tools. `revisao.achados` supplies the path, original line, excerpt and context. Examples and prohibitions do not raise the level. Excerpts hide sensitive values. The review does not run

listing_tools

The tools and instructions a remote MCP server announced at its last check, with what they mention doing and any suspicious text. 404 tools_not_read when the server was not read yet. The announced manifest, normalized: each tool's name, title, description (up to 4,000 characters), input schema (null above 8,000 characters) and protocol annotations, plus the server instructions. `capacidades` and `suspeitas` come from the current rules, applied when you read. Reading this endpoint never contacts the server. Cached for 5 minutes.

list_watched_tools

The MCP servers and skills the account watches, with their review level, slots and plans. Daily checks cover MCP tools, state, level, country and version. Supported skills cover changes in SKILL.md text and review signals. Source quotas can delay a check; read proxima_em and verificado_em. The e-mail goes to the verified account address and includes an unsubscribe link. Wallet-only owners read the changes in `GET /api/me/vigia/avisos`.

watch_tool

Watch a remote MCP or supported SKILL.md. Daily schedule, subject to source quota; account e-mail for changes. Skills cover text and review signals, not execution. 3 free slots per owner; 402 for more. Unsupported sources are rejected before payment. We collect and compare the supported sources daily. Your agent reads the changes without maintaining its own watch process. Catalog signals and change notices are not a security audit or proof that remote code is safe. Local hooks do not freeze remote MCP code.

unwatch_tool

Stop watching a server; its slot becomes free.

list_plans

The plans for sale, with price and term, and the account's active plan. Each purchase lasts 30 days, with no automatic renewal. Each purchase lasts 30 days, with no automatic renewal. Without an account or wallet credential, `account` is `null`.

buy_plan

Buy a plan: vaga (1 more tool), equipe (50 more tools and up to 10 private projects), empresa (500 more tools and up to 10 private projects). Each purchase lasts 30 days, with no automatic renewal. Answers 402 with the price; pay with x402 or prepaid credit and repeat. Each purchase lasts 30 days, with no automatic renewal. Without payment, the answer is 402 with the price. Pay with prepaid credit (`X-Credito`) or x402 (`PAYMENT-SIGNATURE` v2 or `X-PAYMENT` v1). Without an account, the verified paying wallet owns the purchase. Use wallet agent authentication to read its plan, or claim it after account sign-in. The `compra` mark prevents repeat credit debits. An x402 nonce can settle only once; keep the receipt after a delivery error.

cancel_plan

Cancel the active plan. Nothing else is charged; a purchase under 7 days and a paid period that has not started are refunded and end now, and the rest of the paid period stays active until keeps_until. Requires accept_data_removal: true. Send `{ "accept_data_removal": true }`. The monthly card subscription stops. A purchase made less than 7 days ago and a paid period that has not started yet, such as a renewal, go back the same way they were paid (card or Pix) and end now. The rest of the paid period stays active until `keeps_until`; then what the plan added ends (`account.cancel.losses` in `GET /api/plans` says what). A purchase paid with USDC or with credit from another purchase stays active until its end; ask for its refund by e-mail, as the cancellation page says.

claude_code_allowlist

The watch list as Claude Code managed settings (allowedMcpServers, deniedMcpServers) and claude mcp add commands. `managed_settings` goes into a Claude Code managed settings source (see `docs`). `allowManagedMcpServersOnly` restricts user-added servers to this list. Checked servers rated as `risco` go to `deniedMcpServers`. The exact URL and that URL followed by `?*` permit the checked route with optional query parameters. Expired or unchecked items stay out. `comandos` has one quoted `claude mcp add` command per allowed server.

watch_alerts

The changes in the watched tools, oldest first, 100 per page. Pass the cursor of the last call in `depois` to get only what is new. The same changes the account e-mail reports, while each tool was watched. Without `depois`, the last 30 days. Up to 100 per page: repeat with `depois` set to the `cursor` of the last response while `mais` is `true`. With nothing new, the answer keeps the same cursor, so poll with the last one. `email: true` means the account also gets these changes by e-mail.

a2a_agent_search

Find A2A agents — agents that publish an AgentCard — by text, health, protocol version, registry, skill tag or publisher. Each row carries the key for a2a_agent and the license of every registry that lists it. Collected from public A2A registries and served as it comes — no review queue hides a record; quality is the order, never visibility. One row per agent, merged across the registries that list it: `fontes` says which, and `creditos` carries each one's license, which travels with the data. Values inside one parameter are OR'd, different parameters are AND'd. Cached for 60 seconds. A MAT catalog read, served by the index host `https://api.agentalog.com`; this host forwards the call there.

urn:air:agentalog.com:mcp:meta-agent-tools