| name | teacher-profile |
| description | Establish, update, and maintain a single teacher's operating context — their role(s), duties/workload, the handoff & role-interaction map (who they pass work to and receive it from: case manager, AP, counselor, nurse, co-teacher, grade/department team), their school assignment, and personal preferences/defaults — then register it into the shared context as classroom/teacher-scope sop_refs + overrides so every other skill adapts to THIS teacher's reality. Use when a teacher says 'set up my profile', 'this is my role/situation', 'I'm a co-teacher / case manager / department head', 'who handles X at my school', 'remember my preferences', or when another skill needs to know a teacher's duties/handoffs. The profile is a GITIGNORED local store (their data, never committed). Do NOT use for student records (shared/records/), for the school/program directory (canonical-sources/schools/), or for bulk staff directories (shared/staff/, gated) — this is the one teacher's own self-described SOP. |
teacher-profile
What this skill does
Captures one teacher's operating profile through a short setup wizard and keeps it current. A profile
records: the teacher's identity-lite (display name + school, no sensitive PII), their role(s)
(supports multi-role, e.g. "5th-grade teacher + MTSS lead"), duties/workload, a handoff &
role-interaction map (for each recurring handoff: what, to/from whom by role, trigger, cadence), and
preferences/defaults (tone, templates, pacing norms, communication rules). Public sites won't have
the teacher's true day-to-day situation — so teacher-stated facts outrank any crawled inference, and
every field carries provenance + confidence.
The profile is written to a gitignored local store (teacher.local.json under
shared/context/profiles/; shape shown by shared/context/profiles/teacher.example.json, validated by
shared/context/profiles/teacher.schema.json) and registered into the shared context
(shared/context/) as classroom/teacher-scope sop_refs + overrides, so lesson-planner,
family-communication, meeting-classifier, intervention-mtss, special-education-support, and the
records handoffs all adapt without anyone editing a skill.
How it works — the unified pipeline
Follow the shared pipeline in references/method.md
(Request → Routing → Protocol Enforcement → Generation → Validation → Quality Gates → Approval/Certification → Release). The domain work happens in Generation:
Wizard intake → Role/handoff modeling → Context registration (sop_refs/overrides) → Validation.
See references/wizard.md for the interview script and references/profile-model.md for the data model.
- User-truth vs inference — a teacher-stated fact is
provenance: teacher_stated (high confidence);
anything pulled from a public site is provenance: crawled (low/medium) and is confirmed by the
teacher before it shapes behavior. Teacher-stated always wins (shared/context/source-of-truth.md).
- Handoffs are role-based, not person-based — map to roles (case manager, AP, counselor) so the
profile survives staff turnover; an optional person link comes only from the gated
shared/staff/.
- School link — the teacher's school resolves against
canonical-sources/schools/ (MSID), so the profile stays
consistent with the district index.
- Quality — self-check against
references/quality-gates.md, then hand to quality-review.
Privacy & storage (non-negotiable)
- The profile is gitignored (the
*.local.json files under shared/context/profiles/) — the
teacher's own data, opt-in, never committed. A committed teacher.example.json shows the shape with
placeholders only.
- No student PII ever; the teacher's own contact details are minimum-necessary and stay local.
- Updating the profile updates behavior on the next run — no skill edit, no redeploy.
Artifacts
See references/artifact-types.md for the artifact types this skill produces and their specs.
Output: always emit the metadata block
Every artifact ends with the metadata block from protocol-layer/metadata-schema.md, including the
per-dimension quality scores, the decision, and human_review_required: true — outputs are
decision support, not final professional or legal determinations. Use placeholders only; never real
student data.
