Repository navigation
Concepts from docs not included in the skills #1289
Description
Hi, thanks for your work
docs/concepts.md currently state the following:
Good spec content:
- Observable behavior users or downstream systems rely on
- Inputs, outputs, and error conditions
- External constraints (security, privacy, reliability, compatibility)
- Scenarios that can be tested or explicitly validated
Avoid in specs:
- Internal class/function names
- Library or framework choices
- Step-by-step implementation details
- Detailed execution plans (those belong in
design.mdortasks.md)
This information is not included in the skills and instructions Openspec ships with.
Consequently agents don't follow these guidelines unless specifically instructed to.
Activity
clay-good commentedon Jul 1, 2026
Thank you for opening this issue @ao-alexi
I'm adding this into the PR updating agent skills: #1284
docs(skills): embed spec-content guidance in spec-authoring skills (F…
docs(skills): generalize Fission-AI#1289 fix into embedded authoring …
docs(skills): adopt self-contained + guidance-first skills architecture
feat(skills): first build increment — conventions reference + family …
ao-alexi commentedon Jul 8, 2026
thanks @clay-good for your attention to this issue
I see the related PR has been since closed. I assume the direction of OpenSpec is that each team will converge on their own principles and design their config.yaml accordingly?
clay-good commentedon Jul 8, 2026
Hey @ao-alexi,
Thank you for following up. I've opened #1326 to address this.
#1284 was closed for bundling too many changes at once, not as a shift toward per-team configuration and the merged #1300 only carried forward its CLI auto-approval piece.
fix(schemas): include spec content guidance from concepts docs in spe…