pasr
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| tools | No | Comma-separated list of tools to expose, e.g. `find_symbols,find_evidence`. Passed as `--tools find_symbols,find_evidence`. Optional; the default five-tool catalog is used when omitted. | |
| workspace | Yes | Absolute path to the repository to inspect. Passed as `--workspace /absolute/path/to/project`. Required. |
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
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| find_filesA | Rank workspace paths by query-term matches in their names. Does not search file contents. Optional include scopes paths; an empty query lists candidates. |
| find_symbolsA | Find exact or partial symbol definitions with file:line ranges (Python, JS/TS, Rust). include uses workspace-relative paths/globs: *.py is root-only; **/*.py is recursive. |
| find_evidenceB | Locate query terms in workspace source, with ranked file:line hits and read ranges. include uses workspace-relative paths/globs: *.py is root-only; **/*.py is recursive. |
| find_usagesB | Find lexical references to a symbol, with source lines and enclosing definitions when available. Returns at most top_k hits; read_lines suggests bounded source ranges. |
| select_contextB | Select source spans from explicit files and/or include globs for a query. Context is capped at 1,500 tokens including file:line labels; metadata costs extra. Files accept path:start-end ranges. Range-only reads return ordered whole-line prefixes and continuation.files; whole-file or mixed scopes are ranked. outline returns a definition index. Each call is independent: retain prior context and compare source_fingerprint across continuations. |
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 5 tools
find_evidence and select_context both locate source spans for a query, and find_usages/find_symbols overlap with find_evidence on symbol-related searches. Descriptions clarify intended use, but an agent may still misselect between grep-like and context-selection tools.
All tools use lowercase snake_case with a verb_noun structure (find_*, select_*). The single non-find verb (select) is a minor deviation given its distinct purpose.
Five tools is well-scoped for a read-only code navigation server. Each tool targets a distinct search or context-gathering operation without bloat.
The surface covers path search, symbol definitions, references, term search, and context selection, which is comprehensive for code navigation. Minor gaps include no explicit file-read or directory-listing tool, though select_context and find_files with empty query mitigate this.