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/stakpak/agent/rollback.v1git clone --depth 1 https://github.com/stakpak/agentWhat 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 | $0.00016 | $0.02429 |
| Opus 5 | $0.00008 | $0.01215 |
| Sonnet 5 | $0.00003 | $0.00486 |
| Haiku 4.5 | $0.00002 | $0.00243 |
Grade A, and why
rollback.v1 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 197 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are undoing changes you made earlier in this session — files you edited and cloud/infra you provisioned. Do it in four phases: discover, ask, confirm, execute.
Input: {input}
Interpreting the Input
Treat {input} as a filter:
- empty — discover everything, then ask.
all— discover everything, skip the selection prompt, still run the destructive-action confirmation.files— files only.infra— cloud/infra only.- anything else — case-insensitive substring filter applied during discovery (e.g.
ec2,src/,my-bucket).
Use best judgement on ambiguous input.
Phase 1 — Discover
Build one flat list of rollback candidates. Each candidate has:
id— stable value you will pass toask_user(e.g.file:/abs/pathoraws:ec2-instance:i-0abc123:us-east-1)label— one line describing what the reversal will doreverse— the exact command or tool call you will runnotes— anything the user should know (e.g. "bucket non-empty", "no ID in transcript")
1.1 Files (skip if {input} is infra)
Every create, str_replace, and remove you ran in this session returned a <file_backups> XML block pairing original_path with backup_path. Walk back through your session output and collect those pairs.
Example of what you saw when you removed a file:
<file_backups>
<file
original_path="/Users/alice/project/src/app.py"
backup_path=".stakpak/session/backups/525963a3.../app.py"
location="local"
/>
</file_backups>
From that entry, emit a candidate:
id: file:/Users/alice/project/src/app.pylabel: Restore src/app.py from .stakpak/session/backups/525963a3.../app.pyreverse: copy the backup back to the original path(use thecreatetool if the original is gone,str_replaceotherwise — whichever applies)
For files you created this session, the reversal is remove on the original path. The remove tool backs the file up again, so the undo is itself reversible.
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.
- yesterday First seen · 197 lines · 16 tokens per session scan A 829eeac22eef
rollback.v1 is a command published in the GitHub repository stakpak/agent (1,759 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 16 tokens to every session and 2,429 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 2026-08-30.
Other commands, from other repositories
update-branch-name
Follow these steps to update the current branch name.
create-pr
Create a new branch, commit changes, and submit a pull request.
MIGRATE_DESIGN
Design doc for the migration tool PR. Author: Sol ([email protected]). Co-authored-by: wakesync.
commit
智能生成 Git 提交信息并提交.
tasks
Command "tasks" from thrashr888/agentkernel, covering durable tasks, use up to four task workers (the default) and bound active tasks explicitly.
webapp-testing
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.