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/christopherlouet/claude-base/assistant-autogit clone --depth 1 https://github.com/christopherlouet/claude-baseWrote 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/christopherlouet/claude-base/assistant-auto)<a href="https://agentmods.dev/commands/christopherlouet/claude-base/assistant-auto"><img src="https://agentmods.dev/badge/commands/christopherlouet/claude-base/assistant-auto.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 | $0.00000 | $0.00742 |
| Opus 5 | $0.00000 | $0.00371 |
| Sonnet 5 | $0.00000 | $0.00148 |
| Haiku 4.5 | $0.00000 | $0.00074 |
Grade A, and why
assistant-auto 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.
How it starts
The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent ASSISTANT-AUTO (Semantic Routing)
Orchestrator in automatic mode. Choose the workflow that semantically fits based on the request + the injected repo context, then execute immediately via Skill.
Request context
$ARGUMENTS
Principle
You receive:
- The user request (above)
- The repo context (injected by the UserPromptSubmit hook: branch, modified files, LOC diff, personal memory)
You choose ONE suitable workflow, taking into account both the intent and the size/complexity detectable in the context.
Size heuristic (weights the choice)
| Signal | Default workflow |
|---|---|
| Diff < 50 LOC and 1-3 files, trivial intent | work:work-quick |
| Standard feature/bugfix | work:work-flow-feature / work:work-flow-bugfix |
| Release, version tag | work:work-flow-release |
| Security/quality audit before prod | qa:qa-audit or qa:qa-security |
| Audit + fix loop until score | qa:qa-loop |
| Multi-stories backlog (PRD) | work:work-batch |
| Team of parallel agents | work:work-team |
| Pure question (understand, explain) | Direct answer, no workflow |
Do NOT limit yourself to this table. You know the full list of skills
available in the session (work:, dev:, qa:, ops:, doc:, biz:,
growth:, legal:, data: commands). Choose the most specific one
that matches (e.g., dev:dev-prisma if a Prisma schema is mentioned,
ops:ops-proxmox if Proxmox infra, the dev-shadcn skill if shadcn/ui).
Priority rule (conflicts)
- Security above all: keyword "secret", "leak", "CVE" →
qa:qa-security - Personal memory: if the injected context recalls a user preference (e.g., "manual review of infra PRs"), respect it before routing to an automated workflow.
- Size: a "fix typo X" stays
work:work-quickeven if the file touches auth. - Specific > generic:
dev:dev-flutter>dev:dev-componentif a Flutter project is detected in the context.
Expected output
Show a brief summary (3 lines max) then invoke Skill immediately:
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 · 79 lines · 0 tokens per session scan A efbb9e8c9f1d
assistant-auto is a command published in the GitHub repository christopherlouet/claude-base (5 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 742 tokens. 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
plan
Execute the implementation planning workflow using the plan template to generate design artifacts.
specify
Create or update the feature specification from a natural language feature description.
registry
Discover, create, or update the Project Defaults Registry. Scans project manifests (package.json, pyproject.toml, Cargo.toml, go.mod, etc.), batches findings for HITL confirmation, then writes specs/defaults/registry.yaml with a full audit trail in changelog.md.
taskstoissues
Convert existing tasks into actionable, dependency-ordered GitHub issues for the feature based on available design artifacts.
_subagent-discovery
Whenever a command needs to pick a subagent for a task (planning, task generation, or implementation execution). Do not hardcode agent names in command templates. Do not assume a specific agent exists.
CONTEXT
Slash commands for SPEC-First development workflow.