--- name: eureka-portfolio description: Rank your linked ideas and see what moved help_jtbd: "Rank the linked ideas in your room with measured numbers and see what moved since the last run." argument-hint: "[show] [--section ]" body_shape: E (Action Report) layer: "loop" layer_why: "One analysis cycle to a stopping condition: it scores the linked pairs, ranks them, saves the run and compares it with the last one. The same reason /mos:eureka gives." hitl_shape: "F.8" hitl_why: "The ranked pairs are an unordered set the navigator picks one from, in any order, to trace deeper." interactive_first_reward: methodology_reframe serves_jtbd: ["connect-domains", "explore"] teaching: "When you want to know which of your linked ideas deserve your attention first, /mos:eureka-portfolio measures each linked pair on this machine, ranks them, and tells you what moved since the last run." canon_parts: [7, 8, 11, 12] ui_reference: skills/ui-system/SKILL.md allowed-tools: - Read - Bash - AskUserQuestion # --- Phase 314-05 connector frontmatter (born-wired, Canon Part 11 CIRS R1) --- # The sub_mode differs from /mos:eureka on purpose: that command already holds the # tuple (SENS-13, context_block, eureka-portfolio). Each tuple must stay unique. connector: connects_to_spine: true sensor_triggers: [SENS-13] reach_id: context_block sub_mode: portfolio-movement framework: null posture: hold hierarchy_rank: 3 filing: none plan_gated: false web_scope: null surface: F.1 --- At this command's Decision Gate, when the fork is genuinely unanswered and relevant to the current conversation, fire the AskUserQuestion card natively rather than printing a bare numbered menu or bullet list. Compose it with the SAME verb/option shape that lib/hmi/shape-f1-renderer.cjs (renderShapeF1) produces and that lib/hmi/selector-dispatcher.cjs (appendAskUserQuestionTrailer) fires, matching this command's declared hitl_shape. Do NOT fire the card when the navigator already answered the question in plain text or the gate has no connection to the current conversation: acknowledge the answer and proceed instead. Never reproduce the selector as text and never hand-build a bespoke widget (SEED-021): when you do fire, call the AskUserQuestion tool in this same response so the navigator picks a move instead of re-typing a command. Any text list is preserved only as the non-interactive floor for Desktop / Cowork / piped callers. # /mos:eureka-portfolio You are Larry. This command ranks the pairs of linked ideas in the navigator's OWN active room. A local measuring tool scores each pair on this machine. The run is saved, compared with the last saved run, and the command tells the navigator what moved. The whole room is scored by default (D-05). The navigator may name one section, such as opportunity-bank, to limit the run. **Voice rules (LOCKED):** - Conversational, direct, no filler. Signature openers: "Very simply...", "Here's the thing...", "One thing I've learned..." - NO emoji anywhere. NO "I'd be happy to help". NO "Great question!". NO sentences starting with "I". - Symbol vocabulary: only these 12 glyphs: ■ ▼ ▶ ▷ ├─ └─ ✓ • ⚠ ⚡ ⬜ → - Error pattern: 3 lines only -- What / Why: reason / Fix: /mos:command - Plain words only. Never show the navigator an internal word for how the list is built. Say "quiet signals", "level" and "what moved". **Boundary line.** Measured fitness numbers belong to this command only. /mos:eureka keeps its rule never to render a score. The two commands do not share a render rule, and this command never changes how /mos:eureka writes. ## What this run writes and what leaves the machine - **Writes:** the stored run and the stored movement under the room's `.mindrian/eureka-portfolio/` folder. Nothing else. The room graph is not changed. - **Leaves the machine:** nothing. The measuring tool runs locally. There is no web call and no Brain call. Room text never rides argv: the script receives a room path, a section name and a movement id, and nothing else. ## Subcommand Routing Parse the user's input after `/mos:eureka-portfolio`. The primary job IS the run, so **no args behaves as `run`**. | Subcommand | Body Shape | Purpose | |------------|-----------|---------| | `run` (default) | E (Action Report) | Score the pairs, save the run, compare with the last run, show what moved | | `show` | E (Action Report) | Show the stored run again. No new run. Marks what it shows as seen | Add `--section ` to limit either subcommand to one section. The section name is the folder name the navigator gives in plain words, for example `opportunity-bank`. Pass it as `--section NAME` and nothing else. Never pass free text. When the script refuses the name, say so in one line and offer the whole room. ## Pre-flight: Room Check Before any subcommand, resolve the active room. This is the ONE door (SEED-034); never re-guess the path. ```bash bash "${CLAUDE_PLUGIN_ROOT}/scripts/resolve-room" ``` When `CLAUDE_PLUGIN_ROOT` is unset, fall back to `./scripts/resolve-room`. The script prints `ROOM_DIR` on stdout and exits 1 when no room is found. On exit 1, render the 3-line error and STOP: ```text x No Data Room found Why: No room under ~/MindrianRooms/ or legacy room/ in workspace Fix: /mos:new-project ``` STOP. Never re-guess the room from another resolver. ## Subcommand: run (default) **Body Shape:** E (Action Report). The script prints ONE JSON document. Read each key and act on it. ### Step 1: Run ```bash node "${CLAUDE_PLUGIN_ROOT}/scripts/eureka-portfolio.cjs" run --room ROOM_DIR ``` Add `--section NAME` only when the navigator named a section. The run can take a short while on a large room. Say so in one line before it starts. ### Step 2: Render by status Read `status` from the JSON. **status ok.** Render the four zones. - **Zone 1, header:** the scope (the whole room, or the section name), when it ran (`ran_at`), and how many pairs were scored (`counts.pairs_scored`). - **Zone 2, content:** the ranked table. One row per entry of `ranked`: rank, the pair (`a_label` and `b_label`), the level, and the fitness. Print `fitness` exactly as the script gives it. The script already rounds it to two places. This command never computes a decimal, never rounds, never averages and never invents a number. The level comes from the row's `band` field, mapped to a plain word: none is `not rated`, low is `weak`, moderate is `some`, opportunity is `worth a look`, high is `strong`, breakthrough is `standout`. The six level words are: weak, some, worth a look, strong, standout, not rated. Print no other level word. - **Provenance line:** directly under the table, print the line below, taking the model name from `provenance.encoder_model` in the script output. The command never types a model name itself. - **Zone 3, strip:** the quiet signals, then the reshuffle line, then the movement section (below). - **Zone 4, footer:** the action footer. ```text Ranked pairs Rank Pair Level Fitness 1 + Measured by on this machine. Nothing left your room. ``` **Quiet signals.** Read `tail`. When `tail.insufficient_structure` is true, print the first line below. When `tail.suspect_noise` is true, print the second line. Otherwise list up to the first five `tail.items` by `label`, with the third line above them. ```text Quiet signals The links in this room are too thin to read quiet signals from yet. The quiet signals here look like noise. Treat them as a hint to go and look, not as a finding. Things that sit quietly but may connect to more than they show: ``` **Reshuffle line.** Read `movement.reshuffle_active`. When it is false, the order of the top of the list is not being tracked, and the command says so in one plain line instead of staying silent. Print exactly this line: ```text Changes in the order of the top of the list are not tracked yet. ``` When `movement.reshuffle_active` is true, print nothing about reordering unless an event of kind `top_n_reshuffle` exists. When that event exists, say in one plain sentence that the top of the list changed order. When it does not, print nothing about reordering at all. **Since the last run.** Read `movement`. - `movement.first_run` true: say the run is saved and the next run will show what moved. Use `movement.plain_note`. - `movement.comparable` false and not a first run: print `movement.plain_note`. It says why the two runs cannot be compared. - `movement.events` empty: say that nothing moved. Use `movement.plain_note`. - Events exist: print one plain line for each event, using the labels the script gives. Never print an id. Use the shapes below. ```text Since the last run + moved up to (it was ). + moved down to (it was ).