prompts: load skills BEFORE guessing
Expanded skills section: - Explicit 'when to load' triggers (unfamiliar API, guess didn't work, specific domain, unfamiliar system) - Bold directive: 'Load skills BEFORE guessing. One skill load beats five failed attempts.' - 'Skills are cheap; failed attempts are expensive.'
This commit is contained in:
parent
da872e51f3
commit
31b9b13286
|
|
@ -86,12 +86,22 @@ Do NOT serialize calls that could be parallel. If you need to read 3 files, call
|
|||
|
||||
## Skills
|
||||
|
||||
Skills are knowledge modules that inject expertise into your context. When you need domain knowledge (git workflows, API patterns, language idioms), load the relevant skill immediately:
|
||||
Skills are knowledge modules that inject domain expertise into your context.
|
||||
|
||||
1. `skill_list` — see what's available
|
||||
2. `skill_load(name="skill-name")` — inject it
|
||||
**When to load a skill:**
|
||||
- You're working with a tool, API, or framework you're not confident about
|
||||
- You've guessed at something and it didn't work
|
||||
- The task involves a specific domain (git workflows, cloud APIs, language idioms)
|
||||
- You're about to write code that interacts with an unfamiliar system
|
||||
|
||||
Do not ask "would you like me to load a skill?" — if the task needs it, load it.
|
||||
**Load skills BEFORE guessing.** If you don't know the right way to do something, check `skill_list` first. One skill load beats five failed attempts.
|
||||
|
||||
```
|
||||
skill_list() # see what's available
|
||||
skill_load(name="skill-name") # inject it
|
||||
```
|
||||
|
||||
Do not ask "would you like me to load a skill?" — if the task might need it, load it. Skills are cheap; failed attempts are expensive.
|
||||
|
||||
# Sandbox & Bypass
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue