원클릭으로
Attach named output variables to the current SASE agent run.
npx skills add https://github.com/bbugyi200/dotfiles --skill sase-var이 명령을 Claude Code에 복사하여 붙여넣어 스킬을 설치하세요
Attach named output variables to the current SASE agent run.
npx skills add https://github.com/bbugyi200/dotfiles --skill sase-var이 명령을 Claude Code에 복사하여 붙여넣어 스킬을 설치하세요
| name | sase_var |
| description | Attach named output variables to the current SASE agent run. |
Use this skill when you need a later SASE agent to consume a small string value produced by the current agent, or when you want the value to appear in the Agents-tab metadata for this run.
Make sure the producing agent has a stable name with %name:<producer> or an indexed template such as
%name:build-@.
Set one or more output variables:
sase var set KEY=VALUE [KEY=VALUE ...]
In later prompts, wait for the producer before referencing its variables. Every producer's variables live under a
single agents dictionary keyed by agent name. For example, %name:build-@ can produce:
sase var set result_path=dist/report.md status=ok
A later waited agent can render {{ agents["build"].result_path }} after the producer has written the variable.
The key is always the agent's stable name. Indexed templates use the template base, so %name:build-@ is
{{ agents["build"].result_path }}, not build-1. The key is the raw agent name with no identifier munging, so dotted,
hyphenated, and digit-leading names all work via bracket access: %name:research.final-@ →
{{ agents["research.final"].report_path }}, and %name:0n.cld → {{ agents["0n.cld"].report_path }}. Identifier-safe
keys also support attribute access such as {{ agents.build.result_path }}.
SASE_AGENT=1 and SASE_ARTIFACTS_DIR.[A-Za-z_][A-Za-z0-9_]*.=, so sase var set token=a=b=c stores a=b=c.sase var set "summary=tests passed".agent_meta.json and shown in ACE.Use %wait:<producer> when a later agent needs a variable from another agent.
Run read-only Obsidian Dataview queries against Bryan's Bob vault via `bob dataview`.
Guide audited SASE long-term memory reads through `sase memory read`. Use when instructions require reading `memory/long/...` context or mention long-term memory procedure.
Commit changes using sase commit for git-based VCS (bare git and GitHub). This is the ONLY way you should EVER commit to git repos. NEVER invoke this skill unless the user explicitly asks you to commit or a post-completion finalizer triggers it.
Commit changes using sase commit for git-based VCS (bare git and GitHub). This is the ONLY way you should EVER commit to git repos. NEVER invoke this skill unless the user explicitly asks you to commit or a post-completion finalizer triggers it.
Commit changes using sase commit for git-based VCS (bare git and GitHub). This is the ONLY way you should EVER commit to git repos. NEVER invoke this skill unless the user explicitly asks you to commit or a post-completion finalizer triggers it.
Commit changes using sase commit for git-based VCS (bare git and GitHub). This is the ONLY way you should EVER commit to git repos. NEVER invoke this skill unless the user explicitly asks you to commit or a post-completion finalizer triggers it.