AGENT SKILL

recall

Recover context from prior sessions in this repo. Use when the user points at earlier work, asks what happened last time, or resumes an unfinished task.

License 许可
MIT
最近更新
2026-11-12
Stars
★ 2
Forks
0

01 前言 Front Matter

name
recall
description
Recover context from prior sessions in this repo. Use when the user points at earlier work, asks what happened last time, or resumes an unfinished task.
allowed-tools
[ "Bash", "Read" ]

02 资源文件 1

03 SKILL.md 正文

Recall

Recover what the last session decided and left unfinished, then pick the work back up.

Find the session

Transcripts are ~/.claude/projects/<cwd with / as ->/<sessionId>.jsonl. Directory names start with -, so pass absolute paths to ls and grep. Confirm the recorded cwd. The prior session is usually the second newest; order by JSONL timestamps, not file times.

If the session is ambiguous, show each candidate's start, ID, and first prompt. Search this repo by topic before trying other repos.

Read the conversation

Read the dialogue first: the user's typed prompts and the model's prose, in order. Skip tool calls, tool results, file reads, and thinking unless the user asks for them or the dialogue lacks a keyword you expected to find.

jq -r '
  if .type=="user" and (.origin.kind=="human" or .isCompactSummary==true) then
    .message.content
    | if type=="string" then . else (map(select(.type=="text").text) | join("\n")) end
    | "## USER\n\(.)\n"
  elif .type=="assistant" then
    .message.content[] | select(.type=="text") | "## CLAUDE\n\(.text)\n"
  elif .type=="system" and .subtype=="away_summary" then
    "## RECAP\n\(.content)\n"
  else empty end' "$f"

Go deeper on request

For tool calls, thinking, subagent reports, or a specific tool result, read references/records.md for the record types and the queries.

Resume

Summarize the decisions, the work done, and the open thread, then continue the task. Never copy long transcripts into the repo; they can hold secrets.

04 同仓库更多 Skills

notambourine
notambourine/agent-plugins

human-voice

Humanize prose by choosing a surgical voice edit or a structural rewrite. Use for AI-sounding drafts, anti-slop passes, editorial reviews, and requests to preserve an author's voice.

★ 2
notambourine
notambourine/agent-plugins

install-anti-slop

Install and configure the anti-slop Oxlint plugin in a local TypeScript or JavaScript repository. Use whenever a user asks to add anti-slop lint rules, copy the anti-slop plugin, configure opinionated Oxlint rules, or migrate an existing local anti-slop setup.

★ 2
notambourine
notambourine/agent-plugins

improve-codebase-architecture

Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.

★ 2
notambourine
notambourine/agent-plugins

eli5

Explain a topic like I'm a 5 year old. Use when the user types /eli5 <topic> or asks for a dead-simple picture explainer of how something works.

★ 2
notambourine
notambourine/agent-plugins

codebase-design

Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.

★ 2
notambourine
notambourine/agent-plugins

audit-codebase

Read-only whole-codebase audit for simplifications in data structures, state representation, control flow, algorithms, and ownership. Inventories every subsystem as a coverage contract, reviews each one in a bounded subagent, verifies each finding against the repo, then audits its own coverage. Never edits, runs tests, commits, or pushes. Use for "audit this codebase", "find simplifications", or a pre-refactor survey.

★ 2