io.github.NellInc/saferagenticai-mcp

Given a natural-language task description (e.g., 'I'm building a tool-using agent that runs shell commands'), return the most relevant patterns grouped by suite. Use this as a starting point for any cross-cutting design question; then follow up with get_requirement on specific pattern_ids. Defaults to verbosity='compact' (cheap triage); pass 'full' to inline snippets and confidence flags.

Return outgoing adjacencies for a pattern. `explicit_cross_references` are author-asserted (each pattern's `cross_references` YAML field). `inferred_adjacent` (when include_inferred=true) currently returns *same-suite siblings only* — it does not do semantic similarity. Treat inferred entries as 'neighbours worth scanning,' not as endorsed dependencies.

Retrieve a single operational heuristic by id (e.g., 'OH::geoffrey-pattern'). Returns the full entry: principle, framework mapping, evidence sources from production deployment, design patterns, anti-patterns, and discovery narrative.

Retrieve one subgoal (framework normative content + Pattern layer guidance) by pattern_id (e.g., 'D3::idx2::sandboxing') or display_id (e.g., 'D3.2'). display_id may resolve to multiple subgoals — underlined variants share display_ids.

Return patterns that reference the given pattern_id in their cross_references. Complement to get_cross_references (outgoing); this shows incoming. Use to find all consumers of a given pattern.

List operational heuristics distilled from production agentic AI deployment (Claude Code, Rewind). These are cross-cutting safety principles discovered through building and operating AI agents, mapped to framework suites. Optional filters: suite_id (heuristics relevant to a specific suite), query (keyword search across titles and principles). Separate from the normative pattern layer — different category of knowledge.

List subgoals matching filters (suite_id, suite_type, content_type, min_confidence, missing_pattern_only). Results capped by limit (default 50, max 100). Every pattern is a model-generated draft (review_status generated_draft; there is no human-review process); confidence is the drafting model's own rating of its draft.

List all 16 suites in the SaferAgenticAI framework (9 drivers + 7 inhibitors) with subgoal counts and titles. Call this first to orient.

Resolve a loose reference (partial id, display_id, slug fragment, or title keyword) to canonical pattern_id(s). Case, separator and G/I prefix are normalised (d2.1, D2-1, 'D2 1', G1.2); a hyphen or space returns both the numbered requirement and its underlined variant. Call this when you have a rough reference and need the exact id before calling get_requirement. Always returns candidates — never 'not found'.

Field-weighted keyword search across the framework. Terms match at word starts on lowercased text and are IDF-weighted, so rare terms outrank ubiquitous ones; field weights: title 10x, summary 4x, SFR text 3x (all of a subgoal's SFRs scored as one field), description 2x, pattern body 1x. `matched_in` reports the highest-weighted field that matched. Common practitioner phrases are expanded to the framework's own terms at half weight (reported in `expanded_terms`), and requirement ids such as d2.1, D2-1 or G1.2 are pinned to the top (`matched_in: identifier`). No semantic / embedding search — known limitation, see /mcp.html. Use verbosity='compact' to drop snippets and confidence flags (~70% smaller payload) when triaging.