opencode-plugin-creator
Create or revise OpenCode plugins: server entrypoints, hooks, custom tools, lifecycle, installation, and packaging.
v2026.11.13-21cd5682 installs+2 in 7 days39 StarsMITsynced 11 hours ago
Install
No NotHoly CLI yet? Install it first.
$ curl -fsSL https://notholy.com/install.sh | sh$ notholy skills install aeondave/opencode-plugin-creatorWrites to ~/.notholy/skills after checking the signature and the SHA-256.
SignatureSigned, ed25519
Package6afcf1a0…8444920b
6afcf1a0d60b…32838444920b13 files · 9.0 KBScanMedium5 findings
- Executable script shipped with the skillassets/template/src/plugin/guard.ts
- Executable script shipped with the skillassets/template/src/plugin/index.ts
- Executable script shipped with the skillassets/template/test/guard.test.ts
Sourcegithub.com/AeonDave/malskill/tree/main/harness-dev/opencode-plugin-creator@21cd568
On this page
Use opencode-agent-creator for agent definitions, permissions, and task delegation. Keep plugin mechanics here.
Match the host before coding
Record opencode --version and the installed @opencode-ai/plugin version. This skill's template targets OpenCode 1.18.35, tag commit 53d1eabb61e21162157817bf677da0a4ad3332e3.
Choose the API from the target package's exports and types. The released root import uses a Plugin function returning Hooks; its V2 Promise/Effect exports are different APIs. Current V2 documentation can describe fields absent from 1.18.35. Do not mix its examples into a V1 hook object.
For terminal UI behavior, load the official CLI-plugin guide and the target host's @opencode-ai/plugin/tui types. Use its distinct TUI entrypoint/API and package ./tui export; match the guide's imports to the installed version rather than using server hooks for UI work.
Build and verify
- Load references/hooks.md when choosing an entrypoint or interception hook; check its host call site as well as its type.
- Copy assets/template/ into the target project. Set the package name and supported host range; pin the SDK/plugin version used for development.
- Implement only the required hooks. Mutate the provided
outputfields in place. Load references/custom-tools.md when adding model-callable actions. - Load references/patterns.md for initialization, cancellation, logging, state ownership, or teardown.
- Run the template's typecheck and tests in its own directory. Test the returned hooks and custom tool, including the blocked/error path.
- Load references/publishing.md when installing or packaging. Start the matching host in an isolated project, confirm initialization, and exercise the intended behavior. Report unit/typecheck evidence separately from host execution.
- Inspect the package contents before an authorized publication.
Minimal server plugin
import type { Plugin, PluginModule } from "@opencode-ai/plugin"
const Guard: Plugin = async () => ({
"tool.execute.before": async (input, output) => {
if (input.tool === "read" && output.args.filePath === "blocked.txt")
throw new Error("This read is blocked by the project guard.")
},
})
export default { id: "example.guard", server: Guard } satisfies PluginModuleA guard for one tool is scoped to that tool. Set agent/config permissions for access policy; a type declaration alone does not establish that a permission hook runs.
The template exports one explicit server module. Keep helper functions in imported siblings rather than exporting them from a legacy plugin entrypoint.
Files in this version
assets/87.7 KB
template/87.7 KB
test/23.9 KB
- guard.test.ts872 B
- index.test.ts3.0 KB
- .gitignore41 B
- package.json689 B
- README.md961 B
- tsconfig.json410 B
references/412.8 KB
- custom-tools.md2.7 KB
- hooks.md4.0 KB
- patterns.md3.0 KB
- publishing.md3.1 KB
- SKILL.md3.4 KB
Permissions
| Permission | Declared | Found by the scan |
|---|---|---|
| Reads files | Yes | Yes |
| Writes files | No | No |
| Runs shell commands | Yes | Yes |
| Uses the network | Yes | Yes |
| Uses connectors | No | No |
| Drives a browser | No | No |
Scan findings
Reviewed on Nov 13, 2026
- MediumFile · Executable script shipped with the skillassets/template/src/plugin/guard.ts
- MediumFile · Executable script shipped with the skillassets/template/src/plugin/index.ts
- MediumFile · Executable script shipped with the skillassets/template/test/guard.test.ts
- MediumFile · Executable script shipped with the skillassets/template/test/index.test.ts
- InfoLink · Links to outside domains
A clean scan lowers the risk but does not prove a package is safe. Read what you install.
| Version | Published | Commit | SHA-256 | Signature | Permission change |
|---|---|---|---|---|---|
| v2026.11.13-21cd568 | Nov 13, 2026 | 21cd568 | 6afcf1a0 | Signed | – |
More from AeonDave
All from this publisherMore in Reporting and compliance
See the category- skillduty1g/skill5
- rea-ci-tests-auditmorluto/rea-ci-tests-audit4
- tracking-threat-actor-infrastructuremukul975/tracking-threat-actor-infrastructure4
- performing-authenticated-scan-with-openvasmukul975/performing-authenticated-scan-with-openvas4
- slicing-code-contexttrailofbits/slicing-code-context4
- 1337aeondave/13374
Ask anything your scope allows.
Create an account, add credits and ask the first question in the chat or the terminal.
- 1/3
Create the account
Individual or company. You can switch to a company later without losing balance or keys.
IndividualCompany - 2/3
Add credits
By card. Pay per token, no subscription.
$5.00minimum deposit - 3/3
Ask
In the chat, or install the CLI and run notholy login in your terminal.
$ notholy login