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/ashlrai/ashlr-stack/recommendgit clone --depth 1 https://github.com/ashlrai/ashlr-stackWhat 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 | $0.00023 | $0.00276 |
| Opus 5 | $0.00012 | $0.00138 |
| Sonnet 5 | $0.00005 | $0.00055 |
| Haiku 4.5 | $0.00002 | $0.00028 |
Grade A, and why
recommend 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.
What it actually says
Call stack_recommend via the ashlr-stack MCP server with the user's query. Pass save: true so the recipe is persisted, and k: 6 so the top 6 hits come back.
Invocation:
stack_recommend { query: "<user's input>", save: true, k: 6 }
Then render the hits as a short bulleted list, one line per provider:
• <displayName> <category> — <rationale-or-blurb>
After the list, show the saved recipe id on its own line (e.g. Recipe: rec_xxx).
Suggest the next step verbatim:
Run
/stack:apply <recipe-id>to provision these providers and wire their secrets.
If the ashlr-stack MCP server is not configured (tool not available), tell the user to install the plugin with /plugin install ashlr-stack and ensure the stack CLI is on PATH (npm i -g @ashlr/stack).
If the user gave no query, ask them one clarifying question about the goal (e.g. "what are you building?") before calling the tool.
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 · 28 lines · 23 tokens per session scan A 9e1b2d26e867
recommend is a command published in the GitHub repository ashlrai/ashlr-stack (2 stars, last pushed 1mo ago), licensed MIT. It adds 23 tokens to every session and 276 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
pipeline
Run the full delivery pipeline (planner → red-team → coder → smoke-tester → reviewer) against a GitHub issue. Bridges Claude Code's missing auto-handoff button by invoking each subagent in sequence via the Agent tool, validating typed HANDOFF blocks, breaking on repeat failures, and logging every stage to…
10x-plan
Create detailed implementation plans with thorough research and iteration.
10x-research
Research codebase comprehensively using parallel sub-agents.
shadcn-ui-plan
Prompt do wygenerowania wysokopoziomowej architektury UI w .ai/ui-plan.md.
shadcn-view-plan
Prompt do przygotowania szczegółowego planu implementacji widoku (.ai/{view}-view-implementation-plan.md).
10x-implement
Implement technical plans from thoughts/shared/plans with verification.