MCP HubMCP Hub
Volver a habilidades

fungi-identification

pjt222
Actualizado 2 days ago
5 vistas
17
2
17
Ver en GitHub
Otroai

Acerca de

Esta habilidad permite la identificación en campo de hongos mediante el análisis de características morfológicas, esporadas y hábitat. Ayuda a diferenciar especies similares, evaluar toxicidad, y aplica la regla de seguridad de certeza absoluta antes del consumo. Úsela al recolectar, confirmar especies o evaluar hongos desconocidos en una propiedad.

Instalación rápida

Claude Code

Recomendado
Principal
npx skills add pjt222/agent-almanac -a claude-code
Comando PluginAlternativo
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternativo
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/fungi-identification

Copia y pega este comando en Claude Code para instalar esta habilidad

Documentación

Fungi Identification

Find mushroom ID in field. Use shape, spore prints, habitat, season. Safety first, always.

When Use

  • Unknown fungus, need ID
  • Foraging edible mushrooms, confirm species before eat
  • Garden/property fungi: harmful?
  • Build field ID skill with structured observation
  • Tell edible from dangerous look-alike

Inputs

  • Required: Fungus specimen or clear observation in situ
  • Required: Eye for fine detail (cap, gills, stem, base)
  • Optional: Field guide for region
  • Optional: Paper + glass for spore prints
  • Optional: Knife for cross-section
  • Optional: Hand lens (10x) for fine detail

Steps

Step 1: Cardinal Rule

Before any ID work, burn rule into head.

CARDINAL RULE:
If you are not 100% certain of the identification, DO NOT EAT IT.

There is no "universal edibility test" for mushrooms.
Some deadly species taste pleasant.
Some deadly species have delayed symptoms (24-72 hours).
Some deadly species have NO antidote.

The cost of a false positive (eating a misidentified mushroom) is
organ failure and death. The cost of a false negative (skipping an
edible mushroom) is a missed meal.

ALWAYS ERR TOWARD CAUTION.

Got: Cardinal rule internalized before proceeding.

If fail: No failure mode for this step. Rule not internalized → do not proceed to field ID for consumption.

Step 2: Document Habitat

Context narrows ID before touching specimen.

Habitat Recording:
+--------------------+------------------------------------------+
| Factor             | Record                                   |
+--------------------+------------------------------------------+
| Substrate          | Soil, wood (dead/living), dung, leaf      |
|                    | litter, moss, other fungi                |
+--------------------+------------------------------------------+
| Tree association   | What trees are within 10m? (Many fungi    |
|                    | are mycorrhizal with specific tree genera)|
+--------------------+------------------------------------------+
| Moisture           | Dry, damp, wet, waterlogged              |
+--------------------+------------------------------------------+
| Light              | Full shade, dappled, open                |
+--------------------+------------------------------------------+
| Season             | Early spring, late spring, summer, early  |
|                    | autumn, late autumn, winter              |
+--------------------+------------------------------------------+
| Altitude           | Lowland, mid-altitude, montane           |
+--------------------+------------------------------------------+
| Growth pattern     | Solitary, scattered, clustered, ring,    |
|                    | shelf/bracket                            |
+--------------------+------------------------------------------+

Got: Complete habitat record gives context for species ID.

If fail: Habitat unclear (urban garden, mixed plantings)? Record what visible. Incomplete habitat = lower ID confidence — factor into safety check.

Step 3: Examine Morphological Features

Systematic look at specimen.

Morphological Checklist:

CAP (Pileus):
- Shape: convex, flat, concave, conical, umbonate, bell-shaped
- Diameter (measure or estimate)
- Surface: smooth, scaly, fibrous, slimy, dry, cracked
- Colour (note if colour changes with age or moisture)
- Margin: smooth, striate, inrolled, appendiculate (veil remnants)

GILLS / PORES / SPINES (Hymenium):
- Type: gills (lamellae), pores (tubes), spines (teeth), smooth
- Attachment: free, adnexed, adnate, decurrent
- Spacing: crowded, close, distant
- Colour (important — note changes with age)
- Bruising: do gills change colour when damaged?

STEM (Stipe):
- Height and diameter
- Shape: equal, tapered, bulbous, club-shaped
- Surface: smooth, fibrous, scaly, reticulate (netted)
- Interior: solid, hollow, stuffed (pithy center)
- Ring (annulus): present/absent, position, persistent/fragile
- Volva (cup at base): present/absent — ALWAYS check by
  carefully excavating the base (Amanita species have a volva)

FLESH (Context):
- Colour when cut
- Colour change on exposure to air (note time to change)
- Texture: firm, brittle, fibrous, gelatinous
- Smell: mushroomy, anise, radish, flour, chlorine, unpleasant
- Taste: (ONLY if species is confirmed non-deadly by an expert;
  for unknown species, DO NOT taste)

SPORE PRINT:
- Remove the stem; place the cap gill-side down on paper
  (half white, half dark paper to see any colour)
- Cover with a glass or bowl to maintain humidity
- Wait 4-12 hours
- Record spore colour: white, cream, pink, brown, purple-brown,
  black, rust-orange

Got: Full morphological description — all major features.

If fail: Feature not observable (no ring visible, may have been lost)? Record "not observed" not "absent." Distinction matters for ID.

Step 4: ID with Multiple Confirmations

Cross-reference all data vs reference material.

Identification Protocol:
1. Use habitat + season to narrow to likely genera
2. Use cap shape + gill type + spore colour to narrow to species group
3. Check ALL features against the candidate species description
4. Specifically check against dangerous look-alikes:
   - Does this species have a deadly doppelganger?
   - What feature distinguishes the edible from the deadly?
   - Can I see that distinguishing feature clearly?

Confidence Levels:
+----------+---------------------------+---------------------------+
| Level    | Criteria                  | Action                    |
+----------+---------------------------+---------------------------+
| Certain  | All features match; no    | Safe to collect (for      |
|          | look-alike confusion;     | experienced identifiers)  |
|          | experienced with species  |                           |
+----------+---------------------------+---------------------------+
| Probable | Most features match;      | DO NOT eat. Collect for   |
|          | one or two uncertain;     | further study (spore      |
|          | look-alike eliminated     | print, expert review)     |
+----------+---------------------------+---------------------------+
| Possible | Some features match;      | DO NOT eat. Photograph    |
|          | look-alike not fully      | and seek expert opinion   |
|          | eliminated                |                           |
+----------+---------------------------+---------------------------+
| Unknown  | Cannot narrow to species  | DO NOT eat. DO NOT        |
|          |                          | handle extensively        |
+----------+---------------------------+---------------------------+

Got: Species-level ID with explicit confidence + look-alike assessment.

If fail: ID stalls at genus level? OK for learning. For eating, only species-level "Certain" ID acceptable.

Checks

  • Cardinal rule acknowledged before starting
  • Habitat documented before examining specimen
  • All morphological features examined systematically
  • Base excavated to check volva
  • Spore print taken (if time allows)
  • Dangerous look-alikes explicitly checked + eliminated
  • Confidence level honestly assessed
  • Only "Certain" IDs considered for eating

Pitfalls

  • One-feature ID: "Looks like chanterelle" by colour alone. True chanterelles have false gills (ridges), grow from soil near trees, apricot smell. False chanterelles + Jack-o'-lanterns share colour but differ every other feature
  • Skipping base check: No dig = no volva — single most important feature for deadly Amanita (death cap, destroying angel)
  • Trust apps blind: AI mushroom ID apps big error rates for look-alikes. Use as start, never as confirmation
  • "Common = safe": Abundance no tell edibility. Deadly species can be locally abundant
  • Tasting unknowns: Some mycologists taste as diagnostic, needs expert knowledge of safe-to-taste species. Non-experts → no taste unknown fungi
  • Ignoring delayed toxins: Some species (Amanita phalloides) pleasant taste + delayed symptoms. When symptoms appear (24-48h), liver damage severe

See Also

  • mushroom-cultivation — growing known species kills ID risk entire
  • forage-plants — complementary field ID skill; same multi-feature confirmation method

Repositorio GitHub

pjt222/agent-almanac
Ruta: i18n/caveman/skills/fungi-identification
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Habilidades relacionadas

llamaguard

Otro

LlamaGuard es el modelo de Meta de 7-8B parámetros para moderar las entradas y salidas de LLM en seis categorías de seguridad como violencia y discurso de odio. Ofrece una precisión del 94-95% y puede implementarse usando vLLM, Hugging Face o Amazon SageMaker. Utiliza esta skill para integrar fácilmente filtrado de contenido y barreras de seguridad en tus aplicaciones de IA.

Ver habilidad

cost-optimization

Otro

Esta Skill de Claude ayuda a los desarrolladores a optimizar los costes en la nube mediante el ajuste de tamaño de recursos, estrategias de etiquetado y análisis de gastos. Proporciona un marco para reducir los gastos en la nube e implementar una gobernanza de costes en AWS, Azure y GCP. Úsala cuando necesites analizar los costes de infraestructura, ajustar el tamaño de los recursos o cumplir con restricciones presupuestarias.

Ver habilidad

quantizing-models-bitsandbytes

Otro

Esta habilidad cuantiza LLMs a precisión de 8 o 4 bits utilizando bitsandbytes, logrando una reducción de memoria del 50-75% con pérdida mínima de precisión. Es ideal para ejecutar modelos más grandes en memoria GPU limitada o para acelerar la inferencia, admitiendo formatos como INT8, NF4 y FP4. La habilidad se integra con HuggingFace Transformers y permite entrenamiento QLoRA y optimizadores de 8 bits.

Ver habilidad

dispatching-parallel-agents

Otro

Esta Skill de Claude despliega múltiples agentes para investigar y solucionar 3 o más problemas independientes de forma concurrente. Está diseñada para escenarios que involucran fallos no relacionados que pueden resolverse sin estado compartido o dependencias. Su capacidad principal es la resolución paralela de problemas, asignando un agente por cada dominio problemático independiente para maximizar la eficiencia.

Ver habilidad