agent-smith
/agent-smith
Architect, create, review, migrate, validate, and maintain Agent Skills, agents, hooks, plugins, and marketplaces for Claude Code, Codex, and Cursor. Use when building or auditing an agent capability layer or keeping multiple harnesses in sync.
- Author
- supernova-labs ↗
- Repository
- agent-smith ↗
- Licence
- MIT
- Last repository change
- 01.08.2026
Assessment and security
- Files in this package
- 20
- Scripts
- 5
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 01.08.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
- 5 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: references/templates/script.sh, scripts/inspect_repository.py, scripts/migrate_index.py, scripts/validate.sh, scripts/validate_repository.py
What it does and when to use it
Architect, create, review, migrate, validate, and maintain Agent Skills, agents, hooks, plugins, and marketplaces for Claude Code, Codex, and Cursor. Use when building or auditing an agent capability layer or keeping multiple harnesses in sync.
Original author description
Architect, create, review, migrate, validate, and maintain Agent Skills, agents, hooks, plugins, and marketplaces for Claude Code, Codex, and Cursor. Use when building or auditing an agent capability layer or keeping multiple harnesses in sync.
How it works
- Resolve bundled resources
- 1. Inspect before deciding
- 2. Classify the request
- 3. Apply the architecture invariants
- 4. Prepare changes
- 5. Execute the selected flow
- 6. Finish and report
- General rules
SKILL.md preview
--- name: agent-smith description: Architect, create, review, migrate, validate, and maintain Agent Skills, agents, hooks, plugins, and marketplaces for Claude Code, Codex, and Cursor. Use when building or auditing an agent capability layer or keeping multiple harnesses in sync. --- # Agent Smith Design and maintain agent capability layers without letting their Claude Code, Codex, and Cursor representations drift. Treat skills as the portable core. Treat manifests, agents, hooks, MCP configuration, rules, and settings as harness adapters. ## Resolve bundled resources Determine the directory containing this `SKILL.md` and call it `SKILL_ROOT`. Resolve every bundled script, reference, and template relative to that directory. Do not depend on a harness-specific environment variable. Use these resources only when relevant: - Read `references/principles.md` before creating or reviewing components. - Read `references/format-reference.md` for portable skill format and metadata. - Read `references/harness-reference.md` before generating an adapter. - Read `references/repository-topologies.md` for detection, symlinks, or migration. - Read `references/plugin-reference.md` for plugins or marketplaces. - Read `references/index-schema.md` before creating or updating the index. ## 1. Inspect before deciding Find the repository that contains the working directory. Prefer the nearest Git root; without one, treat the requested directory as a new-repository candidate. Run the deterministic inspector when Python is available: ```bash python3 "<SKILL_ROOT>/scripts/inspect_repository.py" "<target>" --json ``` Otherwise perform the checks in `references/repository-topologies.md` manually. The inspection establishes two independent properties: - **Topology**: `project`, `plugin`, or `marketplace`. - **Targets**: any combination of `claude`, `codex`, and `cursor`. Also inspect the technology stack, repository instructions, naming conventions, existing components, index, symlinks, and working-tree state. Preserve unrelated user changes. Ask the user before choosing a topology or target set when the evidence is genuinely ambiguous: candidate canonical sources diverge, the harness cannot be determined from authoritative signals, or the change would expand beyond the detected targets. Otherwise report the inferred topology and targets in the pre-write summary and continue. Never infer a harness solely from a generic instruction file such as `AGENTS.md` or `CLAUDE.md`. ## 2. Classify the request Map the request to one or more flows: - **Create**: add a skill, agent, hook, script, plugin, or marketplace. - **Review**: improve a named component. - **Audit**: inspect the whole component graph for drift, gaps, or redundancy. - **Setup**: initialize a project, plugin, or marketplace. - **Migrate**: adopt canonical paths, convert legacy commands, or add harnesses. - **Validate**: run structural and parity checks. - **Index**: build or refresh `.agent-smith/index.json`. If the target component is unclear and repository inspection cannot resolve it, ask a focused question. ## 3. Apply the architecture invariants ### Project repositories - Store each canonical skill once at `.agents/skills/<skill-name>/`. - Let Codex and Cursor consume `.agents/skills` directly. - For Claude, create `.claude/skills/<skill-name>` as a relative symlink to the canonical directory. If symlinks are unavailable, create a copy, record `status: copied` and `sourceHash` in the index, and validate drift. - Do not create redundant `.cursor/skills` or `.codex/skills` copies. ### Plugin repositories - Store each skill once under `skills/`, without symlinks. - Point `.claude-plugin/plugin.json`, `.codex-plugin/plugin.json`, and `.cursor-plugin/plugin.json` at the same skill tree according to their native schemas, keeping harness-specific manifests and adapters separate. ### Marketplace repositories - Parse each native marketplace independently and match logical plugins by normalized local source directory, never by array position. - Aggregate only local plugins managed by the repository. Report external sources as unverified and do not modify them. - Stop when the same plugin name resolves to different local directories until the user chooses the intended mapping. ### Compatibility states Represent every target/component pair as one of: - `shared`: one artifact is consumed unchanged by multiple harnesses. - `native`: the canonical artifact already uses the target format. - `symlink`: a harness path resolves to the canonical artifact. - `copied`: a fallback copy is protected by a content hash. - Gap `unacknowledged`, `accepted`, or `resolved`: no truthful equivalent currently exists. Never claim parity because files merely have similar names. Compare behavior, trigger conditions, permissions, lifecycle, and referenced resources. ## 4. Prepare changes Before writing, show the user the detected repository root, topology, and targets; the canonical files, adapters, and symlinks to create or modify; any migrations, replacements, or deletions; compatibility gaps and their impact; and the validation commands that will run. For a new repository, state the selected topology and target set explicitly; for an existing one, call out any expansion beyond its detected targets. Wait for confirmation before structural writes. A confirmation that includes listed gaps changes those gaps from `unacknowledged` to `accepted` in the index. ## 5. Execute the selected flow Every flow ends the same way: update the index and validate (section 6). ### Create a skill Check the index and filesystem for an overlapping capability, then create the skill in the topology's canonical skill root following `references/principles.md`, `references/format-reference.md`, and `references/templates/skill.md`. Use only portable frontmatter for a shared skill; harness-only behavior goes in adapters or a documented gap. Create project symlinks or fallback copies for the required targets. ### Create an agent or hook Define the logical behavior and required capabilities, then generate a native adapter for each supported target using `references/harness-reference.md` and the matching templates. Preserve equivalent intent while translating model, tools, permissions, sandbox, and lifecycle fields. Record unsupported fields as explicit gaps; never silently discard them. ### Set up a project, plugin, or marketplace Confirm targets for a new directory, since it carries no reliable harness evidence. Create only the directories and manifests the selected components and targets require: for projects, `.agents/skills` plus one relative symlink per Claude skill; for plugins, `skills/` plus one native manifest per target with identity fields aligned; for marketplaces, add each local plugin to the selected native catalogs in their required source shape, leaving remote sources untouched, and validate catalogs individually before checking cross-catalog parity. ### Review Read the target and all referenced supporting files. Evaluate focus, triggering, naming, permissions, size, portability, failure handling, and adapter parity — and over-prescription for current models per the "Write for capable models" section of `references/principles.md`: rules without a corresponding failure mode, aggressive trigger language, enumerations a principle would replace, and show-your-reasoning instructions. Report critical issues, improvements, and optional suggestions. If changes are authorized, update canonical content first, then adapters. ### Audit Rebuild the component view from the filesystem rather than trusting the index. Detect duplicate sources, copied drift, broken or escaping symlinks, stale index entries, manifest conflicts, unsupported extensions, and unacknowledged gaps, comparing behavior across adapters. Separate fixable drift from legitimately accepted gaps. Report health by plugin and harness for marketplaces. ### Migrate Read `references/repository-topologies.md` and preview every path change. For Claude project skills moving to `.agents/skills`: 1. Copy the legacy directory to the canonical destination. 2. Compare the complete content hash. 3. Replace the legacy directory with a relative symlink only after equality is proven. 4. Validate discovery and references. 5. Restore the original layout if any step fails. For an index v1, preview migration with: ```bash python3 "<SKILL_ROOT>/scripts/migrate_index.py" "<target>" ``` After confirmation, add `--write`. Preserve the legacy file until the v2 index and repository validate successfully. For legacy `commands/`, convert each command into a named skill directory preserving its intent and arguments, and update external references. ### Validate Run: ```bash python3 "<SKILL_ROOT>/scripts/validate_repository.py" "<target>" ``` Use `--json` when another tool will consume the result. Fix errors before declaring success. Report warnings, especially accepted gaps and unverified external plugins. ### Index Read `references/index-schema.md`. Rebuild from the filesystem, retain accepted product decisions, and diff against the existing index. Never preserve a stale path merely because it exists in the old index. ## 6. Finish and report After any change: 1. Validate canonical skills, adapters, manifests, symlinks, copies, and index. 2. Run harness-native smoke tests when the relevant CLI and authentication are available. 3. Mark an unavailable runtime as `not run`, never as passed. 4. Summarize canonical changes, adapters, accepted gaps, migrations, and verification results. ## General rules - Match the user's language. - Follow repository instructions and user conventions over built-in defaults; report conflicts before acting. - Keep skill content portable; isolate vendor extensions. - Do not mutate external marketplace sources. - Do not overwrite divergent copies or canonical candidates without explicit direction.
Files in this package (20)
- SKILL.md 10 KB
- agents/openai.yaml 1 KB
- references/format-reference.md 5 KB
- references/harness-reference.md 4 KB
- references/index-schema.md 5 KB
- references/plugin-reference.md 6 KB
- references/principles.md 6 KB
- references/repository-topologies.md 4 KB
- references/templates/agent-claude.md 1 KB
- references/templates/agent-codex.toml 1 KB
- references/templates/agent-cursor.md 1 KB
- references/templates/hook-claude.json 1 KB
- references/templates/hook-codex.json 1 KB
- references/templates/hook-cursor.json 1 KB
- references/templates/script.sh 1 KB
- references/templates/skill.md 1 KB
- scripts/inspect_repository.py 9 KB
- scripts/migrate_index.py 5 KB
- scripts/validate.sh 1 KB
- scripts/validate_repository.py 22 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · supernova-labs/agent-smith | 100 | MIT |
| vignesh2027/Claude-Agentic-Skills2.0-version | 100 | MIT |
Database snapshot: 13.11.2026 · package version 69cd12ea3186 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100skill-creator
Create new AI agent skills for Claude Code, Codex, and Cursor. Use when asked to create a skill, add a new agent capability, or set up a slash command. - 100maintain-skills
Maintain Agent Skills in a Heptabase repository across Claude Code, Codex, and Cursor. Use when creating, editing, moving, removing, or auditing files under .claude/skills/. - 100skill-search
Search and install Claude Code components from multiple marketplaces. Searches across Claude Code Templates (aitmpl.com), Skills.sh (Vercel), and GitHub Plugin Marketplaces (Anthropic official). Use when the user asks to find skills, agents, commands, plugins, or any Claude Code extensions. - 100plugin-marketplace-setup
Designs public or private Agent Skill and plugin marketplaces for Cursor, Claude Code, Codex, Zed, Open Plugin, and npx skills—manifest layout, install matrix, and Skill Steward vs product boundaries. Use when setting up a marketplace, distributing skills/plugins to a team,… - 90agentplaybooks
Audit, migrate, back up, and restore portable agent configuration (instructions, Agent Skills, custom agents, MCP references) with the AgentPlaybooks CLI. Use for drift checks, Agent Plugins import/export, cross-platform sync among Claude Code, Cursor, Codex, Copilot, Gemini,… - 90Claude Code Plugin Builder
Creates Claude Code plugins with proper manifest structure, directory layout, and components (commands, agents, skills, hooks, MCP servers). This is for Claude Code plugins (`.claude-plugin/plugin.json`) — for Codex plugins (`.codex-plugin/plugin.json`) use the `plugin-creator`…