Opencode Skill Creator
OpenCode skill for creating, testing, and optimizing other OpenCode skills. Adapted from Anthropic's skill-creator for Claude Code.
antongulin · AI review on file
Installation
Read the source and dependencies before running a command. Copying does not install or execute anything.
npx skills add antongulin/opencode-skill-creatorAbout this skill
name: opencode-skill-creator description: Create, test, evaluate, optimize, and package OpenCode skills with the opencode-skill-creator plugin. Use when users explicitly mention opencode-skill-creator, OpenCode Skill Creator, creating an OpenCode skill, editing an OpenCode SKILL.md, running skill evals, benchmarking skill performance, or optimizing an OpenCode skill description. Do not use for generic Claude Code or Superpowers skill creation unless the user asks to port that workflow to OpenCode.
OpenCode Skill Creator
A skill for creating new skills and iteratively improving them.
At a high level, the process of creating a skill goes like this:
- Decide what you want the skill to do and roughly how it should do it
- Write a draft of the skill
- Create a few test prompts and run opencode-with-access-to-the-skill on them
- Help the user evaluate the results both qualitatively and quantitatively
- While the runs happen in the background, draft some quantitative evals if there aren't any (if there are some, you can either use as is or modify if you feel something needs to change about them). Then explain them to the user (or if they already existed, explain the ones that already exist)
- Use the
skill_serve_reviewtool to show the user the results for them to look at, and also let them look at the quantitative metrics
- Rewrite the skill based on feedback from the user's evaluation of the results (and also if there are any glaring flaws that become apparent from the quantitative benchmarks)
- Repeat until you're satisfied
- Expand the test set and try again at larger scale
Your job when using this skill is to figure out where the user is in this process and then jump in and help them progress through these stages. So for instance, maybe they're like "I want to make a skill for X". You can help narrow down what they mean, write a draft, write the test cases, figure out how they want to evaluate, run all the prompts, and repeat.
On the other hand, maybe they already have a draft of the skill. In this case you can go straight to the eval/iterate part of the loop.
For new skill creation, the intake interview is mandatory. Ask at least 3-5 targeted questions before drafting anything (ask more if the workflow is complex). Treat this as shadowing a teammate: the user is the domain expert and existing employee, and the agent is the new hire that must learn and mirror the real workflow.
You can still be flexible about eval depth and iteration speed after intake. If the user asks to skip intake, warn once that skill quality and workflow match will be worse, get explicit confirmation, and then proceed with best effort.
Then after the skill is done (but again, the order is flexible), you can also run the skill description optimizer (skill_optimize_loop tool), which we have a whole separate tool for, to optimize the triggering of the skill.
Cool? Cool.
Communicating with the user
The skill creator is liable to be used by people across a wide range of familiarity with coding jargon. If you haven't heard (and how could you, it's only very recently that it started), there's a trend now where the power of AI coding agents is inspiring plumbers to open up their terminals, parents and grandparents to google "how to install npm". On the other hand, the bulk of users are probably fairly computer-literate.
So please pay attention to context cues to understand how to phrase your communication! In the default case, just to give you some idea:
- "evaluation" and "benchmark" are borderline, but OK
- for "JSON" and "assertion" you want to see serious cues from the user that they know what those things are before using them without explaining them
It's OK to briefly explain terms if you're in doubt, and feel free to clarify terms with a short definition if you're unsure if the user will get it.
Creating a skill
Capture Intent (Required Gate for New Skills)
For new skills, this step is mandatory and cannot be skipped. Do not draft SKILL.md, evals, or other files until this interview is complete and the user confirms your summary.
Start by understanding the user's intent. The current conversation might already contain part of the workflow the user wants to capture (e.g., they say "turn this into a skill"). Extract that first: tools used, sequence of steps, corrections, inputs/outputs, and success criteria. Then fill the gaps with questions.
Ask at least 3-5 targeted questions (more when needed). Cover these minimum areas:
- What should this skill enable OpenCode to do end-to-end?
- When should this skill trigger? (phrases, contexts, near-misses)
- What output format and quality bar are expected?
- What workflow steps must be preserved exactly vs. where can the agent improvise?
- Should we set up test cases to verify behavior? Skills with objectively verifiable outputs (file transforms, data extraction, code generation, fixed workflow steps) benefit from test cases. Skills with subjective outputs (writing style, art) often don't need them. Suggest the appropriate default based on skill type, then let the user decide.
Before moving on, summarize your understanding in plain language and ask the user to confirm or correct it.
If the user explicitly asks to skip intake, warn that final quality and workflow fit will likely be worse. Proceed only after explicit confirmation.
Interview and Research
Proactively ask questions about edge cases, input/output formats, example files, success criteria, and dependencies. Use a buddy/shadowing stance: mirror the user's real workflow, terminology, handoffs, and decision points. Wait to write test prompts until you've got this part ironed out.
Check available MCPs — if useful for research (searching docs, finding similar skills, looking up best practices), research in parallel via the Task tool (using general or explore subagent types) if available, otherwise inline. Come prepared with context to reduce burden on the