Actor capability bound to a harness. Source:
schemas/ontology/skill.schema.json.
What it is
A Skill is a capability that a harness loads, routed to by keyword triggers.
Unlike an Agent, a Skill is required to declare a harness — the safety axis
is part of its identity.
harness is required — a skill cannot exist without declaring where it is
enforced, which keeps the safety axis attached to every capability.
sla_tier lets the self-improving loop know which skills must not regress
when it proposes changes.
Gotchas for an authoring agent
Two AMBIGUOUS edges point at Skill (produced_by, gate_ref). If you author
a Deployment or Risk referencing a skill, spell out the target so the
reference does not degrade to a guess.
1---2name: wiki3description: Skill4---5# Skill67> Actor capability bound to a harness. Source:8> `schemas/ontology/skill.schema.json`.910## What it is1112A `Skill` is a capability that a harness loads, routed to by keyword triggers.13Unlike an `Agent`, a Skill is required to declare a `harness` — the safety axis14is part of its identity.1516## Fields1718| Field | Required | Type / enum | Notes |19|---|:---:|---|---|20| `id` | ✅ | `^[a-z][a-z0-9-]*$` | kebab-case |21| `harness` | ✅ | `claude · kiro · both` | which harness(es) load this skill |22| `triggers` | | array | keyword triggers routed via `.omao/triggers.json` |23| `steering` | | string | optional Kiro steering fragment path |24| `ontology` | | object | `produces` / `consumes` entity refs |25| `description` | | string | |26| `sla_tier` | | `critical · standard · best-effort` | `critical` = must not regress |2728## References (from graph.json)2930- inverse: `Deployment → Skill` via `produced_by` — **produced_by** (⚠️ AMBIGUOUS, 0.5)31- inverse: `Risk → Skill` via `gate_ref` — **blocks_gate** (⚠️ AMBIGUOUS, 0.5)3233## Why it is shaped this way3435- **`harness` is required** — a skill cannot exist without declaring where it is36 enforced, which keeps the safety axis attached to every capability.37- **`sla_tier`** lets the self-improving loop know which skills must not regress38 when it proposes changes.3940## Gotchas for an authoring agent4142- Two AMBIGUOUS edges point at Skill (`produced_by`, `gate_ref`). If you author43 a `Deployment` or `Risk` referencing a skill, spell out the target so the44 reference does not degrade to a guess.
Run npx skillmds@latest add aws-samples/wiki in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Skill It is listed under Coding & Dev Tools on SkillMD.
This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
aws-samples (@aws-samples) published this skill. Their other Agent Skills are listed on their SkillMD profile.
aws-samples