Skip to content
Development
Command

/session-recap

Summarize recent CHANGELOG.md entries for context restoration. Parses the canonical heading shape `## YYYY-MM-DD — pass-N — summary`. Filters: last N (default 10), YYYY-MM date prefix, or "today". Supports --json for machine-readable output.

From plugin
claude-skills
2636 skills28 agents36 commands1 MCP
> /plugin marketplace add heymegabyte/agent-skills

Command definition

session-recap.md
name: session-recap
description: Summarize recent CHANGELOG.md entries for context restoration. Parses the canonical heading shape `## YYYY-MM-DD — pass-N — summary`. Filters: last N (default 10), YYYY-MM date prefix, or "today". Supports --json for machine-readable output.
argument-hint: "[filter] [--json]"
user-invocable: true

<!-- <SUBAGENT-STOP>: skip this skill when running inside a subagent. Meta-skills must not leak into spawned subagent contexts. Source: obra/superpowers `using-superpowers` skill. --> <SUBAGENT-STOP/>

/session-recap

One-keystroke context restoration. When you open a fresh Claude Code session or hand off context to another agent, `/session-recap` returns the canonical CHANGELOG slice you need.

What it does

1. Reads `CHANGELOG.md` from the current repo root (or `$CHANGELOG` env override). 2. Parses entries delimited by `## YYYY-MM-DD — pass-N — summary` headings. 3. Filters by:

  • `<N>` — last N entries (default 10)
  • `YYYY-MM` — all entries from a month (e.g. `2026-06`)
  • `YYYY-MM-DD` — all entries from a specific date
  • `today` — today's entries only (UTC)

4. Returns heading + first 5 body lines per entry (human mode). 5. `--json` mode: structured `{entries:[{date,pass_id,summary,body_preview}],total}` for machine consumption.

Invocation

bash ~/.agentskills/bin/session-recap.sh             # last 10
bash ~/.agentskills/bin/session-recap.sh 20          # last 20
bash ~/.agentskills/bin/session-recap.sh 2026-06     # all June 2026
bash ~/.agentskills/bin/session-recap.sh today       # UTC today
bash ~/.agentskills/bin/session-recap.sh today --json | jq

Output (human mode)

━━━ ## 2026-06-09 — pass-33 — verification-loop self-application + bin/session-recap.sh

  ### Closes both pass-32 Recs
  - **`rules/verification-loop.md` § Self-application** (new) ...
  - **`bin/session-recap.sh`** (new) — context-restoration helper:
  ...

  total: 1 entry

Output (--json mode)

{
  "entries": [
    {
      "date": "2026-06-09",
      "pass_id": "pass-33",
      "summary": "verification-loop self-application + bin/session-recap.sh",
      "body_preview": ["### Closes both pass-32 Recs", "...", "...", "...", "..."]
    }
  ],
  "total": 1
}

When to invoke

  • New Claude Code session on this repo → run first to get context.
  • Brian asks "what did we do today?" → instant grok.
  • Agent handoff → JSON output feeds the next agent's context.
  • Long pause between sessions → date-range filter shows the gap.

See

  • `bin/session-recap.sh` — implementation
  • `rules/verification-loop.md` § Self-application — companion pattern for skill repos
  • CHANGELOG.md — source of truth
Read more
@heymegabyte
Ships withclaude-skills

Agent Skills — agent-neutral autonomous product-building OS for 32+ AI coding tools. 23 skill categories · 28 agents · 149 reference docs · 37 platform variants. One-line prompts → deployed products on Cloudflare Workers.

2628 agents1 MCP
Get the whole plugin

Stars on GitHub

Momentum

stars in total
26
Star history of heymegabyte/claude-skills on GitHub, a chart by star-history.com

Other commands on claude-skills.

bootstrap-status

bootstrap-status

Command

Report the Claude control-plane environment (auth mode, tooling, connectors, resolved target). Fires when you need to confirm the…

More commands from other plugins.

build
Auto-invoked

build

Command

Implement tasks incrementally — build, test, verify, commit. Add "auto" to run the whole plan in one approved pass.

@addyosmani@addyosmaniInstall Command