Problem / Motivation
When users want to run shell commands inside the universal source folder (~/.agents/skills/) or any agent-specific skills folder (e.g. ~/.claude/skills/) — checking symlink targets with ls -la, running git inside a skill, or simply exploring — they currently have to open a terminal manually and re-type or remember the path. This breaks flow for any hands-on inspection or fix.
Acceptance Criteria
Examples
- A user wants to run
ls -la to inspect symlink targets in an agent's skills folder. One click opens a terminal already in that directory.
- A user spots a broken skill and wants to manually
git pull inside the source directory. The app drops them into a terminal there, no typing required.
- A user is debugging permissions on a single agent's skills folder. The terminal opens preset to that agent's path.
Technical design & implementation plan will be produced at pickup time via /plan-eng-review.
Related: #103 (open in Finder) — sibling feature; likely shares discoverability surface and path-validation logic.
Problem / Motivation
When users want to run shell commands inside the universal source folder (
~/.agents/skills/) or any agent-specific skills folder (e.g.~/.claude/skills/) — checking symlink targets withls -la, runninggitinside a skill, or simply exploring — they currently have to open a terminal manually and re-type or remember the path. This breaks flow for any hands-on inspection or fix.Acceptance Criteria
cdafter launchExamples
ls -lato inspect symlink targets in an agent's skills folder. One click opens a terminal already in that directory.git pullinside the source directory. The app drops them into a terminal there, no typing required.Technical design & implementation plan will be produced at pickup time via
/plan-eng-review.Related: #103 (open in Finder) — sibling feature; likely shares discoverability surface and path-validation logic.