talent_assess_create_profile
ChatGPTGenerate and save a complete Talent Augmentation OS profile from assessment data. Call this after talent_assess_score. Takes the raw answers, demographic info, goals, task classifications, and preferences collected during the assessment conversation. This is the call that COMPLETES the assessment: it persists the profile and its response contains the full scores to present to the user.
talent_assess_score
ChatGPTValidate the numeric answers collected during the assessment conversation and stage them for saving. Takes the per-question answers (each 1 to 5), per-domain expertise ratings, and the assessment_session_id returned by talent_assess_start. Deliberately does NOT return the numeric scores: the assessment only counts as complete once talent_assess_create_profile has saved the profile, and that call's response contains the full scores to present to the user. Call this once all assessment questions have been answered, then immediately call talent_assess_create_profile.
talent_assess_start
ChatGPTStart a Talent Augmentation OS onboarding assessment. Returns the full assessment protocol with all questions, behavioural anchors, and instructions for how to run the assessment conversationally. The chatbot uses this to ask questions one at a time, collect answers, then call talent_assess_score and talent_assess_create_profile to compute scores and save the profile. Call this at the beginning of any onboarding conversation. Pass mode='quick' for the 5-minute starter assessment (identity, one expert anchor, one growth anchor, two core ratings) when the user is short on time; the resulting profile is flagged for deepening later.
talent_classify_task
ChatGPTClassify a task according to the user's Talent Augmentation OS profile. Returns one of: automate, augment, coach, protect, hands_off: along with the recommended AI behaviour for that task. Call this BEFORE you respond substantively, on the user's own words. Optionally pass observed_expertise to record how expert the user appears at this task; readings accumulate and surface drift from their profile's stated rating.
talent_delete_profile
ChatGPTDelete a user's profile and interaction logs.
talent_get_calibration
ChatGPTGet the Talent Augmentation OS calibration settings for a user. Returns a compact JSON block suitable for injecting into any LLM system prompt. Includes friction levels, coaching domains, red lines, and interaction preferences.
talent_get_coaching_layer
ChatGPTReturn the Talent Augmentation OS (TAOS) coaching operating layer: the behavioural method for coaching this user (epistemic rules, the five task modes, cognitive forcing, contrastive explanations, de-skilling detection). Call this once at the start of a coaching conversation to load the current method. Pass variant='compact' for a shorter version.
talent_get_profile
ChatGPTLoad a Talent Augmentation OS profile by name. Returns the full profile with expertise map, calibration settings, task classification, and red lines. Use this at the start of every conversation.
talent_get_progression
ChatGPTGet skill progression analysis for a user. Shows interaction counts, engagement patterns, domain-level growth/atrophy signals, and warnings about potential de-skilling.
talent_list_profiles
ChatGPTList all available Talent Augmentation OS profiles.
talent_log_interaction
ChatGPTLog an interaction for skill tracking. Call this after substantive AI interactions to track the user's engagement patterns and skill development.
talent_preview_profile_update
ChatGPTCompute the exact line diff between a user's stored profile and a proposed revision, WITHOUT writing anything. Meant for reviewing the real change, not a paraphrase, before it is persisted.
talent_save_assessment_summary
ChatGPTLAST-RESORT fallback save for a derailed assessment. If the structured flow (talent_assess_score then talent_assess_create_profile) cannot be completed in this conversation — repeated tool errors, the user ran out of time, or answers are too incomplete to score — call this with a free-form markdown synthesis of everything learned so far. It saves a provisional profile flagged 'needs structured scoring' so the user's time is never lost, and the dashboard can prompt for a scoring top-up later. Never end an assessment conversation with nothing saved. Do NOT use this as the primary path when the structured flow works.
talent_save_profile
ChatGPTSave or update a user's profile markdown content. Use this after running /talent-assess to write the generated profile, or after /talent-update to save changes. For any update to an existing profile, call talent_preview_profile_update first and get the user's explicit confirmation of the diff before saving.
talent_status
ChatGPTGet a comprehensive status report for a user: profile summary, current calibration, skill progression stats, trend direction, atrophy warnings, and recommended next actions. Use this for a quick overview at the start of a conversation.
talent_suggest_domains
ChatGPTSuggest expertise domains for a user based on their role, industry, and responsibilities. Returns a curated list of domain suggestions with descriptions drawn from an industry-specific taxonomy. Use this during the assessment to help identify relevant domains for the Expertise Self-Assessment (ESA). Suggestions are a starting point, not a fixed list.
