10 actions in 7 pieces of software · ranked by words, meaning and measured trust · 542 ms
- 1Syntology: code that randemo or tutorialno passport yet
MCP serverrecord
- code_that_ran_find_code_for_paper
Find the code for an ML paper: its repositories, and any implementation Syntology verified. Give an arXiv id (for example 1706.03762), an arXiv URL, or a paper title. Returns `repos`: the repositories the authors declared or Papers with Code cross-referenced (pointers to GitHub; none of these has be
- code_that_ran_get_code_sample
Fetch one code sample by its identity, with everything recorded against it. Use a code_sha256 from search_verified_code. Answers include the verification level and report, the measured input/output behaviour where it exists (and whether that measurement was degenerate, i.e. the test inputs came back
- code_that_ran_search_verified_code
Browse Syntology's index of implementations it has verified, by method name. Returns each implementation's method name, the paper it comes from (arXiv id and title), its publication year, language and verification level, and its code_sha256, which get_code_sample takes. The licence decision comes wi
- code_that_ran_check_coverage
Ask whether Syntology holds code for a method or a paper, before fetching anything. Give a method name (for example "focal loss"), an arXiv id, or a paper title. Answers verified_code, unverified_code, no_code or unknown_name. A no_code answer asserts absence only for the exact name it resolved; unk
- code_that_ran_find_code_for_paper
MCP servergit+https://github.com/formulahendry/mcp-server-code-runnernpmrecord242 stars3,513 downloads/month
- run-code
Run code snippet and return the result.
- run-code
- 3Claude Code Telegram Kitfound in code, not yet checkedno passport yet
MCP serverproject-tharsis/claude-code-telegram-kit
- record_tool
Present when the tool ran inside a subagent; nested tools remain individually disclosed.
- record_tool
- 4code-runner-mcpfound in code, not yet checkedno passport yet
MCP serverdravidsajinraj-iex/code-runner-mcp
- validate_code
Validate code for security and syntax issues without executing it
- validate_code
- 5Python Code Runnerfound in code, not yet checkedno passport yet
- run_python_code
Run Python code in the current environment.
- run_python_code
APIaws/api-models-awsrecord229 stars
- GetSetupHistory
Returns detailed information for five of the most recent SetupInstanceHttps requests that were ran on the target instance.
- GetSetupHistory
- 7code-modedemo or tutorialno passport yet
MCP serverdesplega-ai/code-mode
- run
Execute a saved script (mode='named'), an inline source string (mode='inline'), or source passed via stdin-equivalent (mode='stdin'). Returns { success, result, logs, autoSaved?, ... }. Successful inline/stdin runs are auto-persisted under `.code-mode/scripts/auto/<slug>.ts` (slug derived from `inte
- run
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.