| name | superhero-semantic-model |
| description | Semantic model for the superhero database (comic book heroes and villains — publishers, physical attributes, superpowers, alignment). Load this skill when the user is querying the superhero database. |
| metadata | {"author":"manual","version":"1.1","last_updated":"2026-02-25T00:00:00Z","learned_patterns":0,"verified_patterns":4} |
Superhero — Semantic Model
Available Resources
| Resource | What it contains | How to load |
|---|
| Extended gotchas | Zero sentinel for height/weight, three colour FK columns, NULL publishers, alignment values | read_skill_resource("superhero-semantic-model", "references/gotchas.md") |
| Metric definitions | Average height/weight by publisher, power score distributions | read_skill_resource("superhero-semantic-model", "references/metrics.md") |
| Query patterns | Heroes by alignment, publisher comparisons, power queries | read_skill_resource("superhero-semantic-model", "references/query_patterns.md") |
When to load resources:
- If aggregating height or weight → load
references/gotchas.md (zero sentinel values skew averages)
- If joining to the
colour table → load references/gotchas.md (three FK columns, all to the same table)
- If counting powers or attribute scores → load
references/metrics.md
- For simple alignment or publisher lookups → the information below is sufficient
Key Concepts
| Term (user says) | Actual meaning | Confidence |
|---|
| "hero" / "villain" | alignment.alignment = 'Good' / alignment.alignment = 'Bad' | verified |
| "neutral" | alignment.alignment = 'Neutral' | verified |
| "unknown alignment" | alignment.alignment = '-' | verified |
| "publisher" | publisher.publisher_name — e.g. 'DC Comics', 'Marvel Comics' | verified |
| "eye colour" | superhero.eye_colour_id → colour.colour | verified |
| "hair colour" | superhero.hair_colour_id → colour.colour | verified |
| "skin colour" | superhero.skin_colour_id → colour.colour | verified |
| "power score" | hero_attribute.attribute_value (0–100) for a specific attribute.attribute_name | verified |
| "no colour" | colour.colour = 'No Colour' — treat as missing/not applicable | verified |
Gotchas
-
height_cm and weight_kg use 0 as the sentinel for missing data —
always filter WHERE height_cm != 0 and WHERE weight_kg != 0 before
aggregating physical measurements.
-
Three colour FK columns all join to the same colour table — you must
alias the table three times: once for eye colour, once for hair colour, once
for skin colour. Joining without aliases causes ambiguity errors.
-
publisher_id is NULL for some heroes — not all heroes have a known
publisher. Use LEFT JOIN publisher and handle NULL in output.
-
alignment.alignment includes '-' for unknown — do not assume the
three string values 'Good', 'Bad', 'Neutral' are exhaustive.
-
hero_power is a junction table — one row per hero-power pair — to
count all powers for a hero, COUNT(*) on hero_power after filtering by
hero_id.
For extended gotchas and SQL examples, call:
read_skill_resource(
skill_name="superhero-semantic-model",
resource_name="references/gotchas.md"
)
Core Join Paths
superhero.alignment_id ──→ alignment.id
superhero.gender_id ──→ gender.id
superhero.race_id ──→ race.id
superhero.publisher_id ──→ publisher.id (nullable)
superhero.eye_colour_id ──→ colour.id (alias: eye_col)
superhero.hair_colour_id ──→ colour.id (alias: hair_col)
superhero.skin_colour_id ──→ colour.id (alias: skin_col)
hero_attribute.hero_id ──→ superhero.id
hero_attribute.attribute_id ─→ attribute.id
hero_power.hero_id ──→ superhero.id
hero_power.power_id ──→ superpower.id
Frequently Asked Queries
For canonical query patterns (heroes by publisher, power counts, attribute
comparisons), call:
read_skill_resource(
skill_name="superhero-semantic-model",
resource_name="references/query_patterns.md"
)
Metric Definitions
For full metric formulas (average attributes, power distributions, hero counts),
call:
read_skill_resource(
skill_name="superhero-semantic-model",
resource_name="references/metrics.md"
)
