Spawns three parallel review subagents over a session transcript to route learnings into skill edits.
npx skills add https://github.com/adjohn/pstack --skill reflect-adjohn
Please help me install this Agent Skill. Skill: reflect Source: https://github.com/adjohn/pstack/tree/main/skills/reflect Command: npx skills add https://github.com/adjohn/pstack --skill reflect-adjohn
High-intent search queries and answers about installing and using this skill.
Invoke the reflect skill after a complex task completes. It locates the active session transcript, spawns three parallel reviewer subagents, synthesizes their findings, and presents approved edits to your existing skills for confirmation.
Each reviewer applies a different lens: judgment surfaces durable principles from mistakes and corrections, tooling captures concrete commands and technical facts, and divergent finds blind spots and second-order effects the others miss.
No. The synthesizer produces an Accepted, Rejected, and Backlog list, and the parent agent waits for explicit user approval before applying any edit. Backlog items are filed to your team's tracker separately.
It derives the transcript directory from the current working directory under ~/.claude/projects/<slug>/, then matches the JSONL file whose first user message contains the conversation's opening prompt. It never globs across other project directories.
Skip it when the session was trivial, off-topic, or already covered by an existing skill the agent followed correctly. One-off events and implementation details that drift, like SHAs or version numbers, are not durable learnings.
slbugCapture solved Ruby, Rails, or Grape problems as structured markdown knowledge documents.
firstsun-devCreate, evaluate, and iteratively improve AI agent skills with benchmarked testing.
cursorConverts recurring corrections into lint rules, runtime checks, and automation scripts.
deerawanExecute small edits and configuration changes directly without plan files.
