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.
git clone --depth 1 https://github.com/ashlrai/ashlr-stackWrote 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/ashlrai/ashlr-stack/apply)<a href="https://agentmods.dev/commands/ashlrai/ashlr-stack/apply"><img src="https://agentmods.dev/badge/commands/ashlrai/ashlr-stack/apply/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/ashlrai/ashlr-stack/apply"><img src="https://agentmods.dev/badge/commands/ashlrai/ashlr-stack/apply.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00025 | $0.00558 |
| Opus 5.5 | $0.00010 | $0.00223 |
| Sonnet 5.5 | $0.00005 | $0.00112 |
| Haiku 4.5 | $0.00003 | $0.00056 |
Grade A, and why
apply 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 today.
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
Call stack_apply via the ashlr-stack MCP server with the recipe id the user provided.
Invocation:
stack_apply { recipe_id: "<user's input>", no_wire: false }
What stack_apply does:
- If
.stack.tomlis absent, attempts config-only init with--noInteractive --noProvision. - Runs the CLI provider setup loop for recipe entries one at a time; adapters create or attach resources according to their implementation. On partial failure, the CLI can attempt cleanup of added local config, MCP entries and named vault slots. Headless cleanup can keep secrets; this is not guaranteed upstream-resource deletion. The CLI
--noRollbackflag disables this multi-service cleanup attempt. - After at least one successful setup, reads secret names from
phantom list --json(never values) and reports which declared secrets are in the vault and which are missing. It never writes to the vault;stack addorphantom add <NAME>stores real values. If the vault can't be listed, those providers are reported as skipped. - Writes optional commented webhook stub files for supported recipe providers; the app must implement and verify their handlers.
- Attempts credential, MCP and split-config writes during setup. These effects are not atomic and failures can leave partial state. The final doctor call is informational and does not establish application acceptance.
If the user requests skipping optional wiring (e.g. /stack:apply rec_xxx --noWire), call with no_wire: true. The wrapper forwards the CLI flag --noWire; provider setup, credential storage and config writes still occur. This is not a dry run. Confirm the requested recipe and provider effects before calling.
Read the returned text, including failures and skips, and summarize:
- Which providers were added
- Any providers that were skipped and why (already present, missing API credentials, etc.)
- The
secrets in vault,missingandwebhookscounts apply printed, thenot in Phantom:names, and any providers phantom-wire skipped
Then suggest the next step verbatim:
Run
/stack:doctorto inspect the implemented checks; verify unchecked services and application credentials separately.
If Phantom setup is missing or refused, report the boundary and point to https://phm.dev/secrets for the installed-version setup instructions. Do not initialize a vault or grant permissions automatically. If the recipe id is missing, suggest /stack:recommend <query> to save a new recipe. The MCP server must run in the intended project directory with its Stack CLI and runtime available.
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.
- today Changed · +1 lines · -2 tokens per session 5f6b95504cf0
- 1mo ago First seen · 31 lines · 27 tokens per session scan A 3b5da420fb04
apply is a command published in the GitHub repository ashlrai/ashlr-stack (2 stars, last pushed today), licensed MIT. It adds 25 tokens to every session and 558 once invoked, about $0.0001 per session on Opus 5.5. 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
cost-optimize
Reduce cloud costs across AWS, Azure, and GCP through rightsizing, reserved and spot capacity, storage tuning, and cost monitoring.
data-pipeline
Design and implement batch and streaming data pipelines with ingestion, orchestration, dbt transformations, data quality checks, and monitoring.
deploy-docs
Validate and prepare documentation for GitHub Pages deployment.
setup
Set up Duel Agents routing with your Duel API key.
monitoring-setup.template
This prompt was authored for Claude-style slash workflows. In Codex runtime, adapt tool calls as follows.
tunnel-start.template
This prompt was authored for Claude-style slash workflows. In Codex runtime, adapt tool calls as follows.