Skip to content

SKILL_INDEX.md

Latest commit

 

History

History
65 lines (50 loc) · 7.52 KB

SKILL_INDEX.md

File metadata and controls

65 lines (50 loc) · 7.52 KB
Raw

AKOÚŌ Skill Index

A quick-reference manifest of all portable listening skills in this repository.

Version Notes

v0.9 adds attributable listening passes, listening provenance and corpus lineage, addressable route decisions and coded silence, explicit plural-listening versus ear-swarm declarations, plus corpus-listening and /corpus; it also makes clear that supplied text is not heard sound. v0.8 adds accountable listening context: position, apertures, auditory scales, sources, participants, authority, revision, and honest absence become validated data on every current hearing. v0.7 adds the sovereignty layer: the sovereign-listening mode, the /covenant command, and the listening-covenant schema (schemas/covenant.schema.json). v0.6 instruments the system for host apps with the manifest, presets, memory-lineage listening, apparatus/listener/memory declarations, and budget-aware routing plans.

Meta-Skills

Skill Description
akouo-router Meta-router that analyzes the listening situation, evidence type, risk level, conceptual frame, and assigns primary, secondary, and corrective listening modes. Use this first before any specialized mode.
reference-layer Conceptual mapping layer that turns a sonic object or prior listening output into concepts, sonic methodologies, authors or traditions, research routes, questions, cautions, and adjacent modes. Used by /reference, /study, and /method.

Listening Modes

Skill Description When To Use
signal-inspection-listening Technical signal inspection ear. Audio files, waveforms, spectrograms, metadata, dynamics, frequency, clipping, noise, codecs, compression, machine-listening outputs.
acoulogical-object-listening Perceptual object ear for sound-as-auditum. Sound design, experimental music, Foley, synthetic sounds, ambiguous sources, acousmatic situations, texture-first listening.
embodied-affective-listening Body and affect ear. Bass/club music, noise, drone, ambient, ASMR, sirens, sonic weapons, accessibility, atmospheres of dread or euphoria.
transductive-media-listening Mediation and transduction ear. Sonification, sensors, AI audio, voice cloning, ASR, neural codecs, compression artifacts, platform-mediated recordings.
forensic-archival-listening Evidentiary and archival ear. Recordings of violence, testimony, surveillance, protest recordings, archives, damaged tapes, legal or custody contexts.
ecological-posthuman-listening More-than-human and ecological ear. Field recordings, weather, hydrophones, animal sound, acoustic ecology, urban ecology, soundwalks, habitat and infrastructure listening.
critical-political-listening Critical and political ear. AI audio platforms, surveillance, colonial archives, acoustic justice, sonic policing, marginalized voices, platform critique.
musical-aesthetic-listening Musical, aesthetic, and sound-design ear. Music, rhythm, pulse, meter, tempo, pitch, harmony, interval, melody, contour, timbre, texture, form, production aesthetics, sound-design utility, poetic usefulness, and genre/cultural caution.
symbolic-fictional-listening Speculative and symbolic ear. Sonic fiction, myth, ritual, hauntology, game audio, film worlds, ambient fantasy, creative worldbuilding.
audiovisual-scenic-listening Sound-image-text-scene ear. Video, film, games, UI sound, captions, subtitles, synchronization, offscreen sound, diegesis, audiovisual phrasing, and sound-versus-vision cautions.
voice-speech-listening Voice, speech, and transcript ear. Spoken audio, transcripts, ASR, TTS, voice agents, voice cloning, podcasts, radio, diarization, intelligibility, consent, and identity caution.
accessibility-normative-listening Accessibility and hearing-norm ear. Captions, transcripts, haptics, deaf and hard-of-hearing access, sensory variation, fatigue, masking, alerts, assistive technology, and implied listener audits.
material-event-listening Material, vibration, and event ear. Resonance, vibration, duration, flux, feedback, low frequencies, installation sound, propagation, loudspeakers, room coupling, and sonic process.
memory-lineage-listening Sound-memory and lineage ear. Stored listening records, akousma/akousmata stores, recurrence, kinship, lineage, series over time, archive comparison, and registering listenings into a store.
sovereign-listening Sovereignty ear that listens under a covenant. Community protocols, consent-bound recording, the right to silence and opacity, retention rules, habitat protocols, listening without identification, and auditing hearings against a declared covenant.
corpus-listening Computational-inheritance and corpus-provenance ear. Training, fine-tuning, retrieval, annotation, provider disclosure, licensing, labor, consent, jurisdiction, opt-out, and unknown lineage.

Schema Bundles

Each skill folder includes a references/ directory containing the JSON schemas needed for strict output formatting:

  • listening-output.schema.json — shared output structure for all listening modes
  • listening-context.schema.json — accountable declaration of position, apertures, scale, sources, participants, authority, revision, and honest absence
  • claim-taxonomy.schema.json — epistemic claim categories (heard, measured, inferred, interpreted, speculative, undetermined)
  • listening-pass.schema.json — one attributable listening at one moment
  • listening-provenance.schema.json — evidence sources, cuts, corpora, and optional source-of-voice disclosure
  • route-decision.schema.json — accountable gate outcomes, including coded silence
  • ensemble.schema.json — explicit plural-listening or ear-swarm declaration
  • router-output.schema.json — router-specific output structure (bundled with akouo-router only)
  • routing-plan.schema.json — expanded agent handoff route structure (bundled with akouo-router only)
  • reference-map.schema.json — conceptual mapping output structure (bundled with reference-layer only)

Compatibility

All skills follow the standard skill format:

  • YAML frontmatter with name and description for agent triggering
  • Markdown body with structured instructions
  • Bundled JSON schemas for standalone use

Compatible with any LLM agent framework that supports skill injection (OpenCode, Claude Code, Gemini, etc.).

Using This Index

  • Humans: Browse the table above and click through to individual SKILL.md files.
  • Agents: Load SKILL.md from any skill folder as a system prompt. Schemas are in references/.
  • Developers: Preserve the manifest and bundled-schema conventions, then run ./scripts/validate-release.sh after adding or modifying a skill.