Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/sourikduttanyu/astral/checkpointgit clone --depth 1 https://github.com/sourikduttanyu/astralWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/sourikduttanyu/astral/checkpoint)<a href="https://agentmods.dev/commands/sourikduttanyu/astral/checkpoint"><img src="https://agentmods.dev/badge/commands/sourikduttanyu/astral/checkpoint.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00020 | $0.00334 |
| Opus 5 | $0.00010 | $0.00167 |
| Sonnet 5 | $0.00004 | $0.00067 |
| Haiku 4.5 | $0.00002 | $0.00033 |
Grade A, and why
checkpoint scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Astral checkpoint (hybrid flow). Goal: shed already-finished work from context without losing it — durable record + keep the live thread.
Note: Claude Code has no native partial/verbatim compaction. The only real
selectivity comes from /compact's instruction text. This command writes a
durable summary AND generates that steered /compact line.
-
Scan the session. List each COMPLETED, self-contained work-unit (done + verified, not needed for active work) as short bullets.
-
Use AskUserQuestion (multiSelect) so the user picks which units to shed.
-
For the chosen units, write a concise resume-ready summary (key decisions, file paths, commands, outcomes) to
.astral/checkpoint-<UTC-timestamp>.md. Append a new section if the file already exists. -
Identify what must STAY: the active task and any still-relevant open files.
-
Output a ready-to-run steered compaction line for the user to paste, e.g.:
/compact Keep: <active task> and open files <...>. Drop: completed work — <units>, now saved in .astral/checkpoint-<ts>.md. -
Tell the user: running that
/compactsheds the done work but keeps the thread; the.mdis the durable backup. Do NOT auto-run it — Claude cannot invoke slash commands; the user runs it.
Keep it tight.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 28 lines · 20 tokens per session scan A 8f5ec97bf70f
checkpoint is a command published in the GitHub repository sourikduttanyu/astral (1 stars, last pushed 2mo ago), licensed MIT. It adds 20 tokens to every session and 334 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
recall-loop
Periodic memory digest — designed for use with Claude Code's /loop scheduler. Shows recent memories and pending to-dos in a compact, scannable format.
context-stats
Show context window usage, active tier, and system stats.
resume
Auto-detect and resume any interrupted Plan Cascade task. Detects mega-plan, hybrid-worktree, or hybrid-auto context and routes to the appropriate resume command.
smartcompact
Pin important turns before compacting — pins survive compaction via SessionStart hook reinjection.
implement
Multi-agent advisory squad workflow for implementing code changes: classify → score risk → select agents → planner → Gate 1 (plan approval) → advisory squad → Gate 2 (Blocker halt) → implementer → consolidator → verdict. Auto-detects depth (quick/normal/deep) from risk + file count; pass --quick or --deep to override.…
digest
What happened in a given week, verbatim.