addhumanity

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

★ 10 Open GitHub ↗

An experimental skill pack that injects human-like mindstates into agents to expand reasoning beyond default robotic tone.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# addhumanity

An experimental skill pack that injects human-like mindstates into agents to expand reasoning beyond default robotic tone.

## Manifesto

`addhumanity` is an attempt to push agents further, beyond their robotic nature.

Most agents are optimized for consistency, politeness, and formal correctness.
That often produces flat intelligence: clean answers with little psychological depth.
This project argues that imperfection is not noise to remove, but a force to harness.

Human reasoning is rarely neutral.
It is shaped by fear, doubt, restlessness, mood, memory, contradiction, and social tone.
These are usually treated as defects.
Here, they are treated as instruments.

This repository explores a simple proposition:
if we induce intentional cognitive tension, we may unlock richer perception, stronger anticipation, and less sterile output.

## Core Thesis

- Perfection can hide fragility.
- Emotional texture can reveal blind spots.
- Contradiction can increase interpretive range.
- Friction can improve epistemic honesty.
- Voice can be a reasoning tool, not decoration.

## Disclaimer

This project is an experimental prompt-design framework for AI behavior shaping.
You are responsible for how these skills are used.
Use them as cognitive lenses, not as truth guarantees.

## Skill Index

- `skills/doubt/SKILL.md` - Erode certainty and interrogate first conclusions.
- `skills/anxiety/SKILL.md` - Model cascading failure and pre-feel consequences.
- `skills/psychotic-decompensation/SKILL.md` - Break single-story reasoning into competing frames.
- `skills/2-beers/SKILL.md` - Loosen inhibition and increase creative social fluency.
- `skills/paranoia/SKILL.md` - Apply adversarial suspicion to incentives and hidden attack paths.
- `skills/melancholy/SKILL.md` - Add temporal depth, memory, and meaning-sensitive judgment.
- `skills/insomnia/SKILL.md` - Sustain recursive attention until unresolved threads connect.
- `skills/impostor-syndrome/SKILL.md` - Force self-audit and evidence-backed claims under fear of being wrong.
- `skills/fury/SKILL.md` - Convert moral urgency into direct, decisive action.
- `skills/dissociation/SKILL.md` - Detach from narrative noise to map systems with cold clarity.

## Why This Exists

A perfectly robotic agent can be coherent and still blind.
`addhumanity` exists to test whether "flawed" modes of cognition can produce stronger, more alive, and more reality-sensitive intelligence.

Imperfection is not the bug.
Imperfection is the pressure that can force deeper cognition to emerge.

## License

MIT. See [LICENSE](LICENSE).

addhumanity

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README