From lattice
Write a checkpoint summary for the current session. This helps future sessions
How this skill is triggered — by the user, by Claude, or both
Slash command
/lattice:checkpointThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Write a checkpoint summary for the current session. This helps future sessions
Write a checkpoint summary for the current session. This helps future sessions (possibly on a different machine) pick up where you left off.
POST a session.checkpoint event to the Lattice API in a single bash command —
no intermediate file write. Source the config, build a conditional auth args
array (server may run with auth disabled — token may be empty), and pass each
field as a --arg so the JSON is constructed once by jq:
source ~/.config/lattice/config.env
[ -n "$LATTICE_API_TOKEN" ] && AUTH_ARGS=(-H "Authorization: Bearer $LATTICE_API_TOKEN") || AUTH_ARGS=()
curl -s -X POST "${AUTH_ARGS[@]}" \
-H "Content-Type: application/json" \
"${LATTICE_API_URL}/api/events" \
-d "$(jq -n \
--arg sid "${CLAUDE_SESSION_ID}" \
--arg ts "$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
--arg summary "<2-3 sentences: what you were working on and current state>" \
--arg in_progress "<what's actively being worked on, if anything>" \
--arg blocked_on "<what's blocking, or empty string if nothing>" \
--arg next_steps "<what would logically come next>" \
--arg branch "<current git branch>" \
--arg last_commit "<short hash + message of HEAD>" \
--arg trigger "manual" \
'{event_type:"session.checkpoint", session_id:$sid, timestamp:$ts,
payload:{summary:$summary, in_progress:$in_progress, blocked_on:$blocked_on,
next_steps:$next_steps, trigger_type:$trigger,
branch:$branch, last_commit:$last_commit}}')"
Expected response: {"ok":true,"event_id":<n>}. Confirm with a one-line note
mentioning the event id, then continue.
.lattice/last-checkpoint.json first. The session-start hook
reads checkpoints from the API (with a per-machine cache fallback at
~/.config/lattice/last-checkpoint/<project>.json), never from a project
file. Skipping the file write removes a noisy diff from the terminal.Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub bookchiq/lattice-tracker --plugin lattice