ROBINS I V2 MCP vs Codescene MCP Server
A side-by-side look at the ROBINS I V2 MCP and Codescene MCP Server MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose ROBINS I V2 MCP for developer tools work as a local process; choose Codescene MCP Server for developer tools as a local process. You can also run both side by side in one mcpServers block.
Which MCP server should you choose?
Choose ROBINS I V2 MCP when
- You need dedicated capabilities in the Developer Tools domain.
- You prefer a local stdio subprocess transport.
- You want tools like
get_spec, parse_document, parse_pmcid.
Choose Codescene MCP Server when
- You need dedicated capabilities in the Developer Tools domain.
- You prefer a local stdio subprocess transport.
- You want tools like
analyze_change_set, code_health_refactoring_business_case, code_health_review.
Feature & specification comparison
Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Developer Tools | Developer Tools |
| Quality signal | 55/100 · Good | 65/100 · Great |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Local subprocess (stdio) |
| Auth | Not declared | Not declared |
| Pricing | Not declared | Not declared |
| Required env vars | None listed | None listed |
| Verified owner | Community listing | Community listing |
| GitHub stars | 0 | 72 |
| Directory activity | 1 views 0 installs 0 upvotes | 3 views 0 installs 0 upvotes |
| Tools | 9 | 28 |
| Full listing | View ROBINS I V2 MCP | View Codescene MCP Server |
Tools & capabilities
ROBINS I V2 MCP tools 9
get_specReturn the encoded ROBINS-I V2 specification: 13 preliminaries, 40 signalling questions across 6 domains (domain 1 in two variants), the response vocabularies, the evidence policy, and the algorithm fingerprint. OPTIONAL in the assessment flow — introspection only. assess_result already carries the rubric for the domain in hand, so calling both duplicates the payload. detail='compact' (default) gives question ids, own-words labels, response options and evidence modes; detail='full' adds every intent and assessor note. The descriptions are this implementation's own wording, NOT the published signalling-question text, which is not redistributable. The published question IDs are what make the output interoperable.parse_documentPRIMARY entry point: parse a study report into a bundle with character-offset, source-tagged section spans. Every quote you later cite is resolved against THIS text, and every claim of absence is searched in it. PASS `citation=`: the full bibliographic reference in APA style. It appears on the rendered assessment so the study being judged is unambiguous. `document` is EITHER the raw text OR a file path — but the path must be readable on the SERVER host. If your files live on your own filesystem, paste the TEXT instead, or use parse_pmcid for an open-access PMCID. A path-looking string the server cannot find raises rather than being ingested as its own text. PASS THE SUPPLEMENT WHEN YOU HAVE IT. `supplements` is a list of server-readable paths merged as supplementary material. This matters more for ROBINS-I than for a reporting audit: the target-trial specification that settles C1-C4, and the analysis detail domains 1 and 4 turn on, routinely live only in the appendix. Without it those questions read NI when the answer was merely in a file nobody passed. supplement_status defaults to 'user_provided' when supplements are given; pass 'none_exists' to assert the article has none. Returns the section map, the text hash that keys later calls, and a cue survey — hit counts for the 15 evidence patterns, so you can see before reading where each domain's evidence lives and which absences already look real.parse_pmcidCONVENIENCE entry point: retrieve an open-access article from Europe PMC by PMCID and parse it, main text plus PMC-hosted supplements, merged into one source-tagged bundle. Use it when you have no file in hand, or to auto-fetch an open-access paper's supplement; for a manuscript you were given, use parse_document. supplement_status is 'retrieved' when a supplement was obtained, else 'not_retrieved' — a supplement may still exist on the publisher site, so absence of retrieval is never proof of absence. Raises if no open-access full text is available. An APA-style citation is built from the article's JATS metadata; pass citation= to override it.set_prespecified_confoundersSupply P1 — the confounding factors this REVIEW judges important for the intervention-outcome relationship, listed before any study is assessed. THIS IS BLOCKING. Domain 1 will not be scored without it. Question 1.1 asks whether all IMPORTANT confounding factors were controlled for, and 'important' is defined by this list, not by the paper's covariate table. Substituting the paper's own list would let the study grade its own confounding control, which is exactly the judgement ROBINS-I asks the reviewer to make independently. 'Important' means adjustment would be expected to change the estimate meaningfully; factors with only very weak associations are excluded. You MAY propose a list from domain knowledge or a DAG — that is useful and is what this parameter is for — but a proposed list is not a ratified one. Leave `ratified_by` empty unless a human in this conversation has actually reviewed and accepted the list; the assessment then carries P1 in its ratification queue and is explicitly not final until they do. Do not sign it off on their behalf. Scoped by `review_id` because P1 belongs to the review, not to one study: the same list applies to every result assessed under it.specify_resultName the ONE numerical result being assessed, and settle C4. Required before any domain can be scored. ROBINS-I assesses a single effect estimate, not a paper. A study reporting three outcomes across two analyses yields six assessments; give each its own `result_id` and run them separately. `accounts_for_deviations` IS C4, and it is the highest-leverage input here: it swaps domain 1's entire question set and algorithm, so it cannot be deferred until domain 1 is reached. 'no_itt' — the analysis does NOT account for switches between the compared strategies or other protocol deviations. It targets the effect of assignment. Domain 1 variant A (baseline confounding only). 'yes_pp' — the analysis DOES account for them, by censoring, follow-up partitioning, or a g-method. It targets the effect of sustained receipt. Domain 1 variant B (baseline AND time-varying confounding). Judge this on what the ANALYSIS DOES, not on the label the authors give their estimand. A paper whose protocol table says 'observational analogue of the per-protocol effect' but which never censors at deviation is 'no_itt'. Getting this wrong means answering five questions that do not apply. `result_assessed` is A1: the estimate with its precision. `outcome` is A3. `result_location` (A2) is where it appears and why it was chosen. `target_trial` is C1-C3 as a dict of labelled strings — eligible participants, intervention strategy, comparator strategy, and any note on the estimand. `information_sources` is D1: what you actually read. An NI answer is only defensible relative to what was searched. `b1`/`b2`/`b3` are the section B screening answers (Y/PY/PN/N). b1: was any attempt made to control confounding in this result? b2 (asked only if b1 is PN/N): is the potential for confounding great enough that an unadjusted result should not be considered further? b3: is the outcome measurement method unsuitable for the outcome it is meant to capture? Y/PY on b2 or b3 sends the result straight to critical and no domain is scored.assess_resultReturn the assessment scaffold for ONE domain: the questions actually in play, their own-words intent, the response vocabulary each accepts, what evidence each answer must carry, and the cue searches already run against this bundle with their hits. THE SCAFFOLD IS PER DOMAIN BY DESIGN. Most signalling questions are unreachable on any given path — 24 of 41 were never reached on the reference assessment — and which of domain 1's two question sets exists at all is decided by C4. There is no flat 41-question rubric to fetch, and asking for one would mean answering questions the algorithm discards. domain=0 (default) returns the overview: the preliminaries, what is settled, which domains are done, and which are outstanding. domain=1..6 returns that domain's scaffold. Domain 1 is refused until P1 is set. Work a domain at a time: read the cues to find where the evidence is, answer only the questions the algorithm reaches, then submit_answers. You do not state a judgement — the server computes it from your answers.submit_answersSubmit one domain's signalling-question answers and get back the COMPUTED domain judgement, or (domain=0) finalize the assessment. Each element of `answers` is one question: `question`, `response`, `evidence_mode`, `rationale`, plus the evidence that mode requires — `quotes` for manuscript_positive, `search_cue` or `search_terms` for manuscript_absent, `prior_ref` for reviewer_prior. Answer only the questions the algorithm reaches; omit the rest rather than answering NA. `support` is your narrative for the domain as a whole and appears in the report. Three things are enforced here, and all three are the point of the tool: every quote is resolved to character offsets in the ingested bundle and an unresolvable one is REJECTED with the nearest text found; every claim of absence is backed by a search THE SERVER runs, so it is reproducible; and the judgement is computed by the published algorithm from your answers, never asserted by you. If the traversal reaches a question you did not answer, the call returns status='incomplete' naming it — supply it and call again. `override_judgement` sets the domain judgement against the algorithm and REQUIRES `override_justification`. Overrides cannot hide: the report shows both values and the override enters the ratification queue. domain=0 finalizes: it computes the overall judgement (worst of the six domains by default), assembles the stamped assessment and returns the rendered report inline. `overall_escalate` applies the tool's permitted escalation — several moderates to serious, or several seriouses to critical — and requires a justification. The returned `report` is THE ARTIFACT OF RECORD: present that, not a table of your own, and repeat its ratification queue if it is non-empty, because an assessment with unratified items is not final.render_reportRe-render a finalized assessment as a self-contained HTML page — the meta panel, the six-domain summary strip, the per-domain judgement with its algorithm trail and the evidence behind every answer, styled to the Black Swan Causal Labs identity and carrying the provenance stamp. Returns the HTML in `html` (display it inline or publish it as an artifact) and the same bytes base64-encoded in `content_base64` (decode and save as .html). This is a pure re-render of the artifact submit_answers already stamped — nothing is re-scored. The assessment must have been finalized this session; the cache does not survive a server restart.export_robvisCombine assessment RECORDS from any number of runs into a CSV for **robvis** (McGuinness & Higgins), the standard tool for Cochrane-style risk-of-bias figures. A review of 200 studies is 200 separate runs — each assessment costs a session, and nothing in this server survives between them. So pass `records`: the `record` object each submit_answers(domain=0) returns. They are small, flat and JSON-native, so a whole review's worth fits in one context, and they carry their own provenance so every row stays traceable to a document and an algorithm fingerprint. Omit `records` to use only what was assessed in THIS session (convenient, but session-scoped). READ THE RETURNED `losses` BEFORE PUBLISHING THE FIGURE. It reports records that are not yet ratified, mixed C4 variants, equal weighting, and records built under differing algorithm transcriptions — each of which would make the figure claim more than the assessments support. This is not a column dump, because robvis's ROBINS-I template is ROBINS-I **V1** and V2 is not drop-in compatible: * V1 has SEVEN domains and orders selection of participants BEFORE classification of interventions. V2 has six and swaps that pair. Writing V2's columns out in order loses no data and raises no error — it just prints your classification judgement under the heading "Bias due to selection of participants". layout='robins_i' (the default) places each V2 judgement into its correct V1 SLOT and marks the dropped deviations domain NA. Upload it with tool='ROBINS-I'. * layout='generic' writes six columns headed with V2's own domain names, for tool='Generic'. The headings are then right, but robvis relabels the judgements into ROB1's vocabulary — Moderate becomes "Some concerns", Serious becomes "High". Prefer 'robins_i'. Neither layout can carry 'Low, except for concerns about uncontrolled confounding': robvis reduces every cell to its first initial over a five-fill palette, so it collapses to Low whatever string is written. Say so in the figure caption.
Codescene MCP Server tools 28
analyze_change_setRun a branch-level Code Health review for all files that differ between current HEAD and base_ref. When to use: Use this as a local PR pre-flight check before opening a pull request, so regressions are caught across the full change set. Limitations: - Requires a valid git repository path. - base_ref must exist and be resolvable by git in that repository. - Reviews only files that differ from base_ref. - Output is JSON text from the CLI command. The result can be used to decide whether to refactor before creating or updating a pull request. Returns: A JSON object containing: - quality_gates: the central outcome, summarizing whether the change set passes or fails Code Health thresholds ("passed" or "failed"). - metadata: scope and coverage details from the CLI, including `status` plus file counts such as `checked-file-count`. Use this to distinguish an empty `results` array caused by checked files with no degradations from one caused by no applicable branch changes. - results: an array of objects for each affected file with: - name: the name of the file whose Code Health is impacted. - verdict: "improved", "degraded", "stable", or "unknown". - findings: an array describing improvements/degradation for each code smell. Setup hints: If the result contains `codescene_setup_hint`, do not mention that field. Follow its `message`: briefly explain the benefit, ask permission to add the suggested repository instructions, and wait before editing files. On failure: If this tool returns an error (e.g., the CLI cannot run, or base_ref cannot be resolved by git), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Compare against base_ref="main" for git_repository_path="/repo" and fail the local PR check if any file verdict is "degraded".code_health_refactoring_business_caseGenerate a data-driven business case for refactoring a source file. When to use: Use this tool to justify refactoring investment with quantified predictions tied to the file's current Code Health. Limitations: - Estimates are model-based projections, not guarantees. - Evaluates one file at a time. - Requires an analyzable source file. Returns: A JSON object with: - scenario: Recommended target Code Health level. - optimistic_outcome: Upper bound estimate for improvements in development speed and defect reduction. - pessimistic_outcome: Lower bound estimate for improvements. - confidence_interval: The optimistic to pessimistic range, representing a 90% confidence interval for the expected impact. On failure: If this tool returns an error (e.g., the CLI cannot run, or the file cannot be analyzed), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Call with file_path="/repo/src/service.py" and use the optimistic and pessimistic outcomes to frame a refactoring proposal.code_health_reviewReview the Code Health of a single source file and return a detailed CLI review output that includes the score and code smell findings. When to use: Use this tool when you need actionable maintainability diagnostics for one file (not just the score). Limitations: - Analyzes one file at a time. - Requires a supported source file. - Returns CLI review text, not a normalized JSON schema. Returns: A review string from the CodeScene CLI. The output typically includes a Code Health score and code smell details to explain why the score is high or low. The Code Health scores are interpreted as: * Optimal Code: a Code Health 10.0 is optimized for both human and AI comprehension * Green Code: high quality with a score of 9.0-9.9 * Yellow Code: problematic technical debt with a score of 4.0-8.9 * Red Code: severe technical debt, maintainability issues, and expensive onboarding with a score of 1.0-3.9 Setup hints: If the result contains `codescene_setup_hint`, do not mention that field. Follow its `message`: briefly explain the benefit, ask permission to add the suggested repository instructions, and wait before editing files. On failure: If this tool returns an error (e.g., the CLI cannot run, or Code Health cannot be computed), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Call with file_path="/repo/src/app.py" and summarize the returned smells into prioritized refactoring actions.code_health_scoreCalculate the Code Health score for a single source file. The tool returns one numeric score from 10.0 (optimal) to 1.0 (worst). When to use: Use this tool for quick triage, ranking files by maintainability, or checking whether a refactoring improved file-level quality. Limitations: - Analyzes one file at a time. - Returns only the score summary, not the full smell breakdown. - Requires a supported source file. Returns: A string in the format "Code Health score: <score>". The Code Health scores are interpreted as: * Optimal Code: Code Health 10.0 optimized for human and AI comprehension * Green Code: high quality with a score of 9.0-9.9 * Yellow Code: problematic technical debt with a score of 4.0-8.9 * Red Code: severe technical debt with a score of 1.0-3.9 On failure: If this tool returns an error (e.g., the CLI cannot run, or the score cannot be computed), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Call with file_path="/repo/src/module.py" and compare the score before and after a refactoring.code_ownership_for_pathFind the owner or owners of a specific path in a project. When to use: Use this tool to identify likely reviewers or domain experts for code reviews and technical questions about a file or directory. Limitations: - Requires a valid project_id. - Uses the latest project analysis data available in CodeScene. - If no matching ownership data is found, an empty JSON array is returned. Returns: A list of owners and their paths that they own. The name of the owner who can be responsible for code reviews or inquiries about the file and a link to the CodeScene System Map page filtered by the owner. Explain that this link can be used to see more details about the owner's contributions and interactions within the project. You MUST always show a link after every owner. Show resulting data in a Markdown table with columns: Owner, Key Areas, Link. On failure: If this tool returns an error (e.g., the CodeScene API cannot be reached, or the access token is missing/invalid), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Call with project_id=42 and path="/repo/src/service.py", then present each owner row with its corresponding system-map link.download_skillDownload a single skill to a local directory. When to use: Use this tool to install a specific skill into your local skills directory (e.g., ~/.claude/skills/). Limitations: - By default, refuses to overwrite existing skills. - Set overwrite=true to replace an existing skill. - Creates the destination directory if it does not exist. Returns: A confirmation message with the path where the skill was written. Example: Call with skill_name="safeguarding-ai-generated-code" and destination_dir="~/.claude/skills" to install the skill.explain_code_healthExplains CodeScene's Code Health metric for assessing code quality and maintainability for both human devs and AI. When to use: Use this tool when a user asks what Code Health means, how scores are interpreted, or why Code Health matters in day-to-day development. Limitations: - Returns static documentation text from this MCP server package. - Does not analyze a specific repository or file. Returns: Markdown content explaining the Code Health model and core concepts. Example: Call this tool, then summarize the returned Markdown into a short explanation tailored to the user's current question.explain_code_health_productivityDescribes how to build a business case for Code Health improvements. Covers empirical data on how healthy code lets you ship faster with fewer defects. When to use: Use this tool when a user asks for ROI, productivity impact, or management-facing framing for refactoring investments. Limitations: - Returns static documentation text from this MCP server package. - Does not compute project-specific forecasts. Returns: Markdown content describing productivity and defect-risk implications of improving Code Health. Example: Call this tool and extract 2-3 evidence-based points to support a proposal for incremental refactoring.get_configRead current CodeScene MCP Server configuration values. When to use: Use this tool to discover available configuration keys, inspect effective values, and understand where each value comes from. Limitations: - Returns JSON text only; callers must format it for display. - Sensitive values (tokens) are masked. - Effective values can be overridden by client-provided env vars. When called without a key, lists every available configuration option together with its current effective value, the source of that value (environment variable vs. config file), and a short description. When called with a specific key, returns details for that option only. Sensitive values (tokens) are masked in the output. Returns: A JSON string. When querying a single key, the object has: key, env_var, value, source, description, aliases, and docs_url. When listing all, the object has: config_dir and options (array of the same shape). Use the aliases array to match user intent to the correct key. Present the data clearly and always include docs_url links. Example: Call with key="access_token" to inspect one setting, or call without key to list all configurable options.get_skill_manifestGet the file manifest for a specific skill. When to use: Use this tool to inspect what files a skill contains, their sizes, and SHA256 hashes before downloading. Limitations: - Requires a valid skill name from list_skills. Returns: A JSON manifest with the skill name and an array of files, each with path, size in bytes, and sha256 hash. Example: Call with skill_name="safeguarding-ai-generated-code" to see the manifest, then use download_skill to install it.list_skillsList all available skills embedded in this MCP server. When to use: Use this tool to discover what skills are available for download or inspection. Limitations: - Returns only skills embedded at compile time. - Does not scan external skill directories. Returns: A formatted list of skill names with their descriptions. Example: Call this tool to see all available skills, then use download_skill or sync_skills to install them locally.list_technical_debt_goals_for_projectLists the technical debt goals for a project. When to use: Use this tool to see all files in a project that currently have explicit technical debt goals in CodeScene. Limitations: - Requires a valid project_id. - Returns goal data from the latest available analysis. - Includes only files with non-empty goals. Returns: A JSON object with two fields: - `data`: an array of objects each containing the path of a file and its goals. - `link`: a URL to the CodeScene Code Biomarkers page for the project. Show the goals for each file in a structured format that is easy to read and explain the goal description for each file. It also includes a description, please include that in your output. Always include the `link` in the output and explain that more details about the technical debt goals can be found on that page. On failure: If this tool returns an error (e.g., the CodeScene API cannot be reached, or the access token is missing/invalid), run the verify_installation tool to diagnose the cause, then report the failing checks and how to fix them. Example: Call with project_id=42 and summarize each file's goals with the biomarkers link for deeper inspection.- +16 more on the listing page
Ready-to-paste client configs
Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
ROBINS I V2 MCP
{
"mcpServers": {
"robins-i-v2-mcp": {
"command": "uvx",
"args": [
"robins-i-mcp"
]
}
}
}
Codescene MCP Server
{
"mcpServers": {
"codescene-mcp-server": {
"command": "npx",
"args": [
"-y",
"@codescene/codehealth-mcp"
]
}
}
}
Frequently asked questions
ROBINS I V2 MCP is categorized under Developer Tools and uses a local stdio subprocess. Codescene MCP Server belongs to Developer Tools and uses a local stdio subprocess. Pick ROBINS I V2 MCP for developer tools work and Codescene MCP Server for developer tools.
Copy the mcpServers block from the config section on this page into your client's config (for example claude_desktop_config.json or ~/.cursor/mcp.json), then fully restart the client.
ROBINS I V2 MCP has no declared pricing — check the project's own site. Codescene MCP Server has no declared pricing — check the project's own site. If a server needs environment variables such as API keys, define them under the env key in your client's MCP config.
Yes. MCP clients support multiple servers: add robins-i-v2-mcp and codescene-mcp-server as separate keys in the same mcpServers object, and your assistant will call tools from either one.
On AllMCPs, ROBINS I V2 MCP has a quality signal of 55/100 (Good) with 1 views, 0 installs and 0 GitHub stars. Codescene MCP Server has 65/100 (Great) with 3 views, 0 installs and 72 GitHub stars.