From 4a1c6ee61b0d402c91c3db2c2a55575fdfb17b49 Mon Sep 17 00:00:00 2001 From: Aswin Prakash Thiyagarajan Date: Fri, 28 Aug 2026 15:18:58 -0400 Subject: [PATCH] docs(mcp): drop ask_ollie now that the tool is disabled MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit opik-mcp#170 disabled ask_ollie everywhere behind a kill switch, so the MCP no longer exposes it. This removes it from the MCP server docs: the tool table (now five tools), the self-hosted and Cursor-timeout notes, the example conversation, and the "Ollie & auto-approve" section (auto-approve only ever governed Ollie's mid-stream writes, so it is moot without the tool). run_experiment is untouched — it is still Cloud-only and available. Co-Authored-By: Claude Opus 4.8 --- .../docs-v2/prompt_engineering/mcp-server.mdx | 35 ++++++------------- 1 file changed, 10 insertions(+), 25 deletions(-) diff --git a/apps/opik-documentation/documentation/fern/docs-v2/prompt_engineering/mcp-server.mdx b/apps/opik-documentation/documentation/fern/docs-v2/prompt_engineering/mcp-server.mdx index 2abf9489215..79311449faf 100644 --- a/apps/opik-documentation/documentation/fern/docs-v2/prompt_engineering/mcp-server.mdx +++ b/apps/opik-documentation/documentation/fern/docs-v2/prompt_engineering/mcp-server.mdx @@ -169,7 +169,7 @@ you change something. ### The tools you'll have -Your assistant gets six tools. It picks between them on its own — this is here so +Your assistant gets five tools. It picks between them on its own — this is here so you know what it can reach for: | Tool | What your assistant can do with it | @@ -178,7 +178,6 @@ you know what it can reach for: | `list` | Page through any of those, optionally filtered by name. | | `write` | Log traces and spans, score, comment, save prompt versions, manage test suites and experiments. | | `schema` | Look up the exact payload shape for a write, so it constructs valid ones. | -| `ask_ollie` | Investigate or synthesize across entities via Opik's in-product assistant. | | `run_experiment` | Run an evaluation experiment end to end. | To see a payload shape yourself, ask **"show me the schema for trace.create"** — @@ -401,8 +400,8 @@ connect to a named cloud workspace. **Cursor 60s timeout.** Cursor enforces a hard tool-call timeout that does - not reset on progress notifications. Long `ask_ollie` turns will fail on - Cursor — see [Known client limits](#known-client-limits). + not reset on progress notifications. Long-running tool calls (for example a + large `run_experiment`) can fail on Cursor — see [Known client limits](#known-client-limits). @@ -507,31 +506,17 @@ connect to a named cloud workspace. **Self-hosted Opik.** Add `COMET_URL_OVERRIDE` to the `env` block (and `OPIK_URL` -if Opik lives at a non-default path). `ask_ollie` and `run_experiment` are -available on Comet Cloud only — on self-hosted those calls fail at dispatch; -use `read` / `list` / `write` directly. +if Opik lives at a non-default path). `run_experiment` is available on Comet +Cloud only — on self-hosted that call fails at dispatch; use `read` / `list` / +`write` directly. -## Ollie & auto-approve - -By default, writes that Ollie performs mid-stream (scores, comments, prompt -versions, test-suite items) execute without a per-action confirmation step. -Each auto-approved write is logged as a JSON audit row on the `opik_mcp.audit` -Python logger. - -To require manual confirmation instead, set `OPIK_MCP_AUTO_APPROVE=disabled` in -the server's `env` block. Ollie's confirmation requests then surface as typed -errors that you can re-issue manually. - -`ask_ollie` and `run_experiment` are available on Comet Cloud only — on -self-hosted those calls fail at dispatch; use `read` / `list` / `write` -directly. - ## Known client limits - **Cursor enforces a 60-second hard tool-call timeout** that does not reset on - progress notifications. Long `ask_ollie` turns will fail on Cursor. For - long-running investigations, use Claude Code or VS Code Copilot. + progress notifications. Long-running tool calls (for example a large + `run_experiment`) will fail on Cursor. For long-running investigations, use + Claude Code or VS Code Copilot. ## Example conversation @@ -539,7 +524,7 @@ A typical investigative loop using Claude Code: > **You:** Why did the experiment "gpt-4o-rerank-v3" regress on factuality? > -> **Claude:** *(calls `ask_ollie`)* Three traces failed because the reranker +> **Claude:** *(calls `list` and `read`)* Three traces failed because the reranker > dropped the system message. The remaining 12 traces scored above 0.8… > > **You:** Score the bottom 3 traces 0.2 with reason "dropped system message".