Skills Registry Skill Dropdown v2 Q10Analysis Card Q11Section Confirm Q12Write Gate Q37Handoffs Q38Mode Switch Q48Grant vs Contract Document Rating All Answers
AI · FS_9 Skills Architecture CANON

Skills Registry — 11 skills, 4 groups

Every AI surface in the product runs exactly one skill per context (FS_9 rule). Each skill is a SKILL.md file with a mode key, an injected-context contract, and an entry surface. This registry is the single source of truth for the skill dropdown, the skill bar, and routing. New skills are added here first, then to the dropdown.

01 · Foundation — builds the company working set
02 · Discovery — finds and loads opportunities
03 · Response — analyzes and writes
04 · Quality — gates everything that ships
Structure
.sk-grid — auto-fill minmax(255px,1fr)
.sk-card — bg1, bd2, hover border accent
.sk-key — mono 2xs accent, is the routing mode key
.sk-meta — Context in / Entry / Output rows
.sk-group-h — mono label + 1px rule fill
Rules
Exactly one skill active per context (FS_9)
Status: shipped / build / spec — drives dropdown visibility
spec skills hidden from dropdown behind flag skills.preview
Registry is data-driven: skills.registry.json renders both this grid and the dropdown
Opportunity Response · FS_2 Q10 ANSWERED

Solicitation Analysis Card

Shown on the opportunity detail page after the user clicks Respond. The opportunity_response skill reads the opportunity fields + attached solicitation and renders this card. Four derived stats, a factor list reusing the GO/NO-GO chip language, and a footer that advances to section confirmation. While analyzing, the card shows skeleton rows with status ANALYZING…

Opportunity · NSF STEM K-12 · after “Respond”Q10
Solicitation Analysis
NSF-26-547 · STEM_K12_NOFO.pdf · analyzed in 41s
COMPLETE
Match
High
profile vs criteria
Eligibility
GO
high confidence
Sections found
8
from solicitation
Page / word limits
15p
narrative cap
Qualifies501(c)(3) status confirmed; Mississippi service area matches geographic priority; prior NSF sub-award counts as past performance.
QualifiesRequired partnership letter — Jackson State University MOU on file in Knowledge Center.
RiskIndirect rate agreement not on file. De minimis 10% allowed — confirm before budget narrative.
GapLogic model required as Appendix B — profile has no logic model for the STEM Pathways program. Added to Unmet Requirements.
Next: confirm the 8 sections extracted from §IV.2 “Proposal Preparation Instructions”
DECISION
Q10: The analysis card is a single full-width card on opportunity detail, not a modal. States: analyzing (skeleton rows + pulsing dot) → complete → error (retry). Factor chips reuse the Eligibility Analysis vocabulary (Qualifies / Risk / Gap). Gap items write to the same Unmet Requirements inventory the contract writer reads (Q48) — one inventory, two consumers. The footer CTA is the only path forward; analysis is never skippable when a solicitation is attached.
Structure
.an-card — 680px max, bd2
.an-stats — grid-4, mono values, hi/med color
.an-row + .an-chip — chip-ok / chip-warn / chip-gap
.an-foot — note + Re-run + primary CTA
States & data
analyzing — skeletons, CTA disabled
complete — shown above
error — danger band + Retry
Writes: analysis_result, section_schema[], unmet_requirements[]
Opportunity Response · FS_2 / FS_4 Q11 ANSWERED

Section Confirmation UI

The user confirms the section schema before anything is written. Sections extracted from the solicitation carry a “From your solicitation” tag; the user can rename inline, reorder by drag, toggle include/exclude, and add custom sections (tagged “Added by you”). Confirming locks the schema and opens the write gate (Q12). The FS_4 missing-instructions rule renders the amber gate below — never a silent preset fallback.

Section confirmation · 8 from solicitationQ11
Confirm response sections
Extracted from NSF-26-547 §IV.2. Rename, reorder, or exclude — generation follows this list exactly, in this order.
⠿11 pgFrom your solicitation
⠿22 pgFrom your solicitation
⠿36 pgFrom your solicitation
⠿4—From your solicitation
⠿52 pgFrom your solicitation
⠿62 pgFrom your solicitation
⠿71 pgFrom your solicitation
⠿82 pgFrom your solicitation
⠿9—Added by you
+ Add a section
8 included · 1 excluded · order locked at confirm
Missing-instructions gate · FS_4 ruleQ11
No proposal instructions found
This opportunity has no attached solicitation and no instruction document, so there are no sections to extract. Upload the NOFO, RFP, or proposal preparation instructions to continue — generation never starts from a preset section list.
DECISION
Q11: Section confirmation is a dedicated card, not inline chat. Row anatomy: drag grip · order number · inline-rename input · page/word limit (extracted) · source tag (From your solicitation / Added by you) · include toggle. Excluded rows dim to 45% and are skipped (count as resolved per FS_4). Confirm writes the locked section_schema the writer consumes — section labels and ordering come from this list only, which also answers the Q48 sub-question. The eight-section default list exists only as the analysis fallback suggestion and is never used silently.
Analysis → Generation · FS_2 step 6 Q12 ANSWERED

Write Gate — full dialog copy

The single checkpoint between analysis and generation. It shows exactly what the writer will use, surfaces the one thing worth fixing first (readiness), and sets expectations for the hybrid auto-approve first pass. This copy is final — ship it verbatim.

Write gate · modal over opportunity detailQ12
DECISION
Q12 copy, verbatim: Eyebrow = Eligibility complete → Grant Writing (the FS_9 named transition). Title = “Ready to write your response?” Body = “The Grant Writer will draft N confirmed sections for [opportunity], in order, using everything below. You review and approve each section — nothing is submitted anywhere.” Checklist rows: Sections confirmed · Solicitation attached · Eligibility · Knowledge Center · Profile readiness (warn row links to the fixing skill, e.g. Story Collection). Note block explains hybrid auto-approve + [VERIFY] in two sentences. CTAs: secondary = Not yet — add impact stories (contextual to the warn row); primary = Start writing →. Contract variant swaps eyebrow to Eligibility complete → Contract Response and the body noun to “response volumes.”
Skill Transitions · FS_9 Q37 ANSWERED

Visible Handoffs — “Profile Ready” + carried-context strip

Two handoff patterns, both visible. Intake → anywhere: a full “Profile ready” card replaces the chat when intake confirms, itemizing exactly what was carried forward. Analysis → generation: a persistent context strip pinned at the top of the writer showing what it inherited — dismissible, recoverable from the toolbar.

Handoff A · intake confirmed → Profile Ready cardQ37
✓
Profile ready
The Bean Path is set up for grant pursuit. Here's what every skill now knows about you.
Carried forward to all skills
IdentityThe Bean Path · Jackson, MS501(c)(3) · EIN on file
PathGrantschosen at end of Stage 1
Programs3 programs · 1 logic modelSTEM Pathways · Tech Literacy · Innovation Lab
Certifications2 activeMBE · pending: 8(a)
Documents4 indexedcapability statement · 990 · MOU · bios
Readiness72%Impact stories are the biggest lift → Story Collection
Handoff B · carried-context strip pinned in writerQ37
Carried from analysis
SECTIONS 8 confirmed MATCH High ELIGIBILITY GO SOURCE STEM_K12_NOFO.pdf LIMITS 15 pg narrative
✕
DECISION
Q37: Intake handoff is a summary card, not a transitional screen — it replaces the chat thread in place so there's no dead navigation moment. Six carry rows: Identity, Path, Programs, Certifications, Documents, Readiness; the readiness row names the weakest weight and links the skill that fixes it. CTAs route by path choice (Grants → Find opportunities). Analysis handoff is the pinned .ho-strip in the writer: five key-values (sections, match, eligibility, source file, limits). Dismiss collapses it to a toolbar chip — the context never disappears, only the strip.
Skill Transitions · FS_9 Q38 ANSWERED

Named Mode Switch — leaving a skill mid-flow

From anywhere, the skill dropdown is always reachable (chat panel header, writer toolbar). Switching is named and announced — a switch banner states what's being left, what happens to it, and what's shared. Drafts always autosave; switching is never destructive. The only guard: a generation in flight.

Switch banner states · rendered at top of chat on switchQ38
Grant Writer→Grant Search Draft saved at 62% · resume from Knowledge → Drafts
Grant Intake · Phase 3→Eligibility Analysis Intake progress kept · partial profile is shared context
Story Collection→Grant Writer 2 new stories indexed · available to generation immediately
Guarded switch · generation in flightQ38
DECISION
Q38: One switching mechanism everywhere: the grouped dropdown (chat header) plus a compact skill chip in the writer toolbar that opens the same dropdown. Context model: company profile + Knowledge Center are shared across all skills; each skill keeps an independent chat thread and working state, keyed by (skill, opportunity). Mid-draft switch: autosave, banner with resume path; in-flight generation continues in the background and lands in ready_review without auto-approving. Returning to a skill resumes its thread exactly — no restart, no replay.
Grant Writer · FS_4 / FS_9 Q48 ANSWERED

One Writer Shell — Grant vs Contract surfaces

The writer is one shell, two skills. Layout, state machine (not_started → generating → ready_review → approved + skipped/error), progress math, [VERIFY] footer, autosave, and export are identical. The active skill changes the vocabulary and switches three contract-only surfaces on. Section labels and ordering come from the confirmed schema (Q11) only — never hardcoded per skill.

NSF STEM K-12grant_generation
Executive Summary50 wordsapproved
Statement of Need42 wordsapproved
Project Narrative61 wordsready_review
Goals & Objectivesgenerating
Work Plan & Timelinenot_started
[VERIFY] markers 3
Exec Summarynumber of students reached in pilot year
NarrativeJSU faculty count for co-training
Goalsbaseline STEM participation rate
AZ-DOT I-17 DBEcontract_generation
L.4.1Technical Approachapproved
L.4.2Management Planready_review
L.4.3Past Performancegenerating
L.5Price Volumenot_started
Unmet requirements · GAP 2
L.2.3Bonding letter required over $2M — not on file
M.1.4NBIS certification copies for 2 named inspectors
Labor categoryRateHrsExt
Sr. Bridge Inspector$142/hr1,200$170,400
NBIS Team Lead$168/hr640$107,520
CAD / Reporting$96/hr480$46,080
DECISION
Q48, all four sub-questions: ① The writer UI does not change structurally between skills — same shell, same six-state badges, same progress math. ② Section labels/ordering come from the confirmed schema only; contract sections additionally render a .wr-secref mono prefix (L.4.2 / M.1) when the analysis extracted solicitation numbering. ③ Contract-only surfaces, flag-driven by skill: the GAP Unmet Requirements panel (separate from [VERIFY], per FS_4), section-number references, and the rate table component inside the Price Volume section. Grants show the [VERIFY] panel in the same slot. ④ Mixed grant + contract on one opportunity: not supported in one document. One response draft = one skill. An opportunity may hold two drafts (one per type), listed separately under All Drafts.
Shared shell
State machine, progress %, autosave, Copy/Copy All, Export to Google Docs, reopen guard — identical
.wr-status — st-approved / st-review / st-gen / st-notstarted (+ skipped, error)
Contract flags
skill.flags.gap_inventory → Unmet Requirements panel
skill.flags.section_refs → .wr-secref prefix
skill.flags.rate_table → .rate-tbl in price sections
All three default off for grant_generation
Quality · Rubric-gated NEW SKILL

Document Rating — rubric scorecard

Document Rating requires a rubric — funder evaluation criteria extracted from the solicitation, a saved internal rubric, or one the user defines on the spot. No rubric → the skill asks before scoring, never invents criteria. The scorecard rates each criterion with evidence, and the overall score feeds the writer's per-section rating recompute.

Rubric scorecard · scored against NSF merit criteriadocument_rating
84/ 100
Draft v3 — NSF STEM K-12 response
Rubric: NSF-26-547 merit review criteria · extracted from §VI · 5 criteria, weighted
Intellectual Merit · 30%
Strong partnership model; cite the prior pilot's published outcomes to anchor the claim.
26/30
Broader Impacts · 30%
Underserved-population reach is clear; quantify teacher-training multiplier effect.
25/30
Evaluation Plan · 15%
Indicators map to objectives. Add data-collection cadence and instrument names.
10/15
Organizational Capacity · 15%
Section currently excluded from draft — reviewers will look for it. Re-include or fold into Narrative.
6/15
Budget Justification · 10%
Costs trace to activities; confirm de minimis indirect rate language.
9/10
Scores recompute on edit · history kept per draft version
Rubric sources, in priority
1 · Extracted funder criteria (solicitation §)
2 · Saved org rubrics (Knowledge Center)
3 · User-defined inline (name + weight per criterion)
No rubric → blocking picker, mirrors the Q11 amber gate
Structure
.ru-score — 52px boxed total, ok/warn/danger border by band
.ru-row — criterion + evidence note + bar + n/N
Bands: ≥80% ok · 50–79% warn · <50% danger
“Fix lowest criterion” deep-links the writer section
FS_9 · Closure OVER-COMMUNICATE

Every open question — answered

The full FS_9 open-question list, resolved by the designs above. Hand this table with the page to Cursor; each row maps to a directive in SKILLS_DS_DIRECTIVES.md.

Q#QuestionAnswer (design on this page)Status
Q10Analysis card wireframeSingle card on opportunity detail: header + 4 stats + Qualifies/Risk/Gap factor rows + footer CTA. States analyzing/complete/error. Gap rows feed the shared Unmet Requirements inventory. #q10Answered
Q11Section confirmation UIDedicated confirm card: drag-reorder, inline rename, limits, source tags, include toggles, add-row. Missing-instructions amber gate replaces silent preset fallback. Confirm locks section_schema. #q11Answered
Q12Write-gate dialog full copyFinal copy shipped verbatim in the modal: eyebrow transition name, 5-row checklist, hybrid/[VERIFY] expectation note, contextual secondary CTA. #q12Answered
Q37Visible handoff momentIntake → “Profile ready” summary card (6 carry rows + readiness gap + path-routed CTAs). Analysis → pinned carried-context strip in writer, collapsible to toolbar chip. #q37Answered
Q38Mid-flow skill switchingOne grouped dropdown everywhere + writer toolbar chip. Shared: profile + KC. Independent: per-skill thread keyed (skill, opportunity). Autosave + named switch banner; in-flight generation finishes to ready_review. #q38Answered
Q48Writer UI: grant vs contractOne shell. Schema-driven labels/order only. Contract flags: GAP inventory, L/M section refs, rate table. Mixed grant+contract = two drafts, never one document. #q48Answered