Second Read
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TAVILY_API_KEY | No | Web search key. | |
| SECONDREAD_FAKE | No | Set to 1 for demo mode: no keys, no model calls. | |
| ANTHROPIC_API_KEY | No | Anthropic model key. | |
| SECONDREAD_ENV_FILE | No | Path to a .env file to load; otherwise the server reads .env from its working directory. | |
| SECONDREAD_MAX_WORDS | No | Document size cap (default 20000). | 20000 |
| SECONDREAD_MODEL_MID | No | Override the model used for the mid tier. | |
| SECONDREAD_MODEL_CHEAP | No | Override the model used for the cheap tier. | |
| SECONDREAD_MODEL_STRONG | No | Override the model used for the strong tier. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| review_documentA | Review a document draft with a desk of specialist agents: a fact-checker that cites sources, a devil's advocate that builds the strongest counter-case, a structure reviewer, a copy editor and an originality check. Returns a markdown report (must fix, counter-case, should fix, polish, originality notes, claim ledger). Pass exactly one of |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 1 tool
There is only one tool, so there is no possibility of selecting the wrong one. The description clearly delineates its scope (multi-agent document review) with an explicit mutual-exclusivity rule between `content` and `url`.
The single name `review_document` follows a clear verb_noun snake_case convention and is self-explanatory. With only one tool there is no pattern to cross-check, so consistency is trivially satisfied but unverifiable.
A single tool is defensible for a narrow, single-capability server, but the surface is thin: everything (fact-checking, counter-case, structure, copy edit, originality) is bundled into one opaque call with no way to request a subset of reviewers.
The one tool covers the full review report in a single call, but a job that 'takes a few minutes' has no status/polling or async retrieval path, and there is no follow-up operation (e.g. re-review a revised draft or compare versions). Agents can work around this by re-invoking, but it is a notable gap for a long-running operation.