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/paulbreuler/limps/run-agentgit clone --depth 1 https://github.com/paulbreuler/limpsWhat 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 | $0.01735 |
| Opus 5 | $0.00000 | $0.00868 |
| Sonnet 5 | $0.00000 | $0.00347 |
| Haiku 4.5 | $0.00000 | $0.00173 |
Grade A, and why
run-agent 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 2d 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 — 213 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run Agent
Select the next best agent task from a plan and open it in Cursor, or run a specific agent file.
Invocation
/run-agent [plan-name]
/run-agent --agent [agent-path]
/run-agent --assess [plan-name]
/run-agent --agent [agent-path] --open-only
What This Command Does
-
Selects next best task (if plan name provided) - Uses scoring algorithm based on:
- Dependencies (40%): Features with all dependencies PASS get highest score
- Workload balance (30%): Agents with fewer remaining tasks get higher score
- Priority (30%): Lower feature IDs (earlier in plan) get higher score
-
Assesses agent status - Checks completion state and file organization
-
Validates agent path (when
--agentprovided) - Errors if file does not exist -
Opens agent file in Cursor - Opens the selected or specified agent file with context
-
Displays instructions - Shows quick links and next steps
-
Prints a brief checklist - Extracts Files/Tests sections when present (LLM-focused)
Usage Examples
Select and Run Next Task
/run-agent 0004-datagrid-overhaul
This will:
- Analyze the plan to find the next best agent task
- Display selection with score breakdown
- Verify agent status
- Open the agent file in Cursor
- Show context and instructions
Run Specific Agent File
/run-agent --agent plans/0004-feature-name/agents/004_agent_name.agent.md
Opens the specified agent file directly in Cursor.
To find agents: Use /plan-list-agents <plan-number> to see all available agents with clickable links, then use /run-agent --agent [path] to run the specific one you want.
Assess Agent Status
/run-agent --assess 0004-datagrid-overhaul
Assesses all agents in the plan for completion status and file organization.
Instructions for Claude
When this command is invoked, you must:
- Get next task or parse arguments:
- If plan name provided: Use
get_next_taskMCP tool (server:limps) to get the next best task - If
--agentprovided: Use the specified agent file path - If
--assessprovided: Useget_plan_statusMCP tool (server:limps) to assess plan status
- If plan name provided: Use
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.
- 2d ago First seen · 213 lines · 0 tokens per session scan A 37a438a77819
run-agent is a command published in the GitHub repository paulbreuler/limps (10 stars, last pushed 6mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,735 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-new
Create a new plan using the Universal Planning Framework.
plan-review
Check plan quality against Universal Planning Framework standards.
interview-plan
Interview a plan using framework-aware questions across 3 depth tiers.
plan-refine
Autonomously stress-test and harden a plan through 6 adversarial perspectives.
start
Enter Fathom Mode with a task. Bare /fathom:start shows status only via hook (no LLM).
plan
Plan from the current Fathom session, grounded in a structured intent.