8,072 tools
list_paths
List all of the user's learning paths. Call this on a return visit to re-orient yourself.
mark_step
Update a step's progress status on the user's behalf, optionally attaching a note about their experience.
update_step
Edit a step's title, resource URL, rationale, kind, or time estimate. Fails on user-locked steps.
reorder_steps
Reorder a path's steps. Provide the complete list of step ids in the new order. User-locked steps must stay at their current positions.
record_note
Attach a note to a step on the user's behalf — their reaction to a resource, a confusion, a takeaway. Notes are shown on the step card and i…
get_progress
Read a progress summary for a path: per-step statuses, notes, completion counts, remaining time, and days since last activity. watched_perce…
create_quiz
Create a multiple-choice quiz attached to a step. The USER answers it in the page UI — do not answer it yourself. Set reveal: 'on_step_done'…
get_quiz_results
Read the user's answers to a quiz. Returns pending status until the user submits in the page. Use missed questions to adapt the path.
await_quiz_results
Wait (block) until the user submits a quiz in the page, then return their results — call this right after create_quiz instead of polling. Re…
propose_revision
Propose a set of path changes for the USER to accept or decline in the page UI — use this instead of direct edits when a change is significa…
await_proposal_decision
Wait (block) until the user accepts or declines a proposal in the page — call this right after propose_revision instead of polling. Accepted…
get_proposal_status
Check whether the user accepted or declined a proposal. Accepted ops are already applied — read the path with get_path afterwards. A decline…
highlight_step
Point the user at one step on the page: it scrolls into view with your short message attached. Use when telling the user where to pick up or…
create_flashcards
Attach a flashcard deck to a step for the USER to review in the page — front is the prompt, back is the answer. Good after the user finishes…
get_review_stats
Read the user's flashcard results for a deck: per-card seen/got-it counts and last result. Cards repeatedly marked 'again' signal concepts t…
plan_weeks
Assign steps to calendar weeks (1-based) so the path fits the user's stated hours per week — each step shows its week badge in the UI. Provi…
get_application_status
Get overall application progress: sections completed, open questions, and whether it's ready to submit.
list_open_questions
List every unanswered or unconfirmed required field — the agent's work queue.
list_blocking_items
List everything currently blocking submission.
explain_requirement
Get the plain-language requirement and permission class for one field.
get_section
Get all fields and current values for one section, by section id.
get_provenance_report
Get who filled every field and when — the full provenance trail.
list_evidence
List evidence files attached to this application.
list_review_items
List reviewer comments across all sections.
get_history
Get the full, replayable mutation history for this application — who changed what, in order.
navigate_to_section
Move the active section — affects which section-scoped tools are registered next.
answer_question
Answer a field directly. Refuses with a structured reason for human_only fields (attestations, signatures) and evidence_required fields (use…
propose_answer
Propose a value for an evidence_required field (e.g. transcribed from an attached document). Lands as 'needs your confirmation' — never sets…
validate_section
Check one section for unanswered required fields.
request_human_action
Escalate to the human: flag a field or general issue that needs their attention. Surfaces in the 'needs you' queue.
request_evidence
Ask the human to attach a document supporting an evidence_required field. The agent cannot attach files itself.
add_review_comment
Add a review comment to a section, authored under the caller's own identity.
check_consistency
Cross-reference every answered field for contradictions: overlapping dates, an expired passport, income figures that don't add up.
summarize_application
Get a plain-language summary of application progress: what's done, what's outstanding, and any consistency issues found.
read_evidence
Read an attached evidence file and extract a plain-text summary of what it appears to contain, so the agent can decide what to propose_answe…
navigate_magic_brain
Open a named Magic Brain page. Use the remote research tools to retrieve card or market data.
search_cards
Search Magic Brain's public card catalogue by name or text, with optional set, color, rarity, and type filters. Use this to discover card ID…
get_card
Get public catalogue details for one Magic card by ID, including printing and set metadata when available. Does not return ownership, watchl…
get_latest_prices
Get the latest available public market prices for up to 100 card IDs in one read-only request. Price records preserve the API's source, obse…
get_price_history
Get bounded public historical price observations for one card and date range. Use daily or weekly intervals; the maximum range is 366 days. …