A tool that ranks candidate dysfunctions against observed behaviour invites exactly one misuse — someone diagnosing themselves or another person — and this one names the cases where self-report is unreliable rather than leaving it implicit. The coverage statistics do the same job from another angle, exposing per-confidence breakdowns instead of presenting the database as uniformly authoritative. Read what this is: a structured reference over one book's classification, useful for research, writing and understanding a framework's internal structure. It is not a clinical instrument, it does not follow DSM or ICD, the severity rubrics carry an early version number, and nothing here substitutes for a qualified assessment of a real person.
A reference database over a book's classification of behavioural dysfunctions, organised on axes, with tools for narrowing candidates, reading severity rubrics, and seeing which patterns confuse with which.
- list_compromised_self_report is a transparency tool: it returns which dysfunctions cannot be reliably self-diagnosed, which is an unusually direct statement of the framework's own limits.
- get_differential_map returns everything that confuses with a given pattern, in both directions — what it is mistaken for and what is mistaken for it.
- differential_diagnosis ranks candidate patterns against observed behaviours and returns scores rather than a single answer.
- get_dysfunction returns one pattern's full entry, optionally filtered to specific modalities, and get_probe returns the elicitation content for a modality.
- score_severity returns the rubric applied to observations, with the version stated.
- suggest_intervention returns tiered first-line and second-line options.
- list_axes and list_dysfunctions cover the structure, with the tool text noting which axes are canonical to the source book.
- review_stats reports coverage — total entries and the breakdown by axis, confidence and self-report reliability — so how complete the data is can be read rather than assumed.
- resolve_id canonicalises a partial id, slug or name into the real one.
Nothing to supply. The content follows a specific published framework rather than a clinical standard such as DSM or ICD.
One command — uvx psychopathia-mcp