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 skills add lindoelio/spec-driven-steroids --skill long-running-work-planninggit clone --depth 1 https://github.com/lindoelio/spec-driven-steroidsWrote 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/skills/lindoelio/spec-driven-steroids/long-running-work-planning)<a href="https://agentmods.dev/skills/lindoelio/spec-driven-steroids/long-running-work-planning"><img src="https://agentmods.dev/badge/skills/lindoelio/spec-driven-steroids/long-running-work-planning/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/skills/lindoelio/spec-driven-steroids/long-running-work-planning"><img src="https://agentmods.dev/badge/skills/lindoelio/spec-driven-steroids/long-running-work-planning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 82 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00067 | $0.01284 |
| Opus 5 | $0.00034 | $0.00642 |
| Sonnet 5 | $0.00013 | $0.00257 |
| Haiku 4.5 | $0.00007 | $0.00128 |
Grade A, and why
long-running-work-planning 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 10d 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 — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Long-Running Work Planning
Make extended agent work durable, resumable, and observable. The primary goal is not deeper private reasoning; it is keeping progress outside the model context through files, task states, checkpoints, and verification.
Core Principle
For heavy work, never rely on memory or a long internal chain of thought as the source of truth.
Use durable artifacts instead:
requirements.md,design.md, andtasks.mdfor spec-driven work- task status markers in
tasks.md - small verified batches
- explicit checkpoint summaries in the working artifact or final response
- validation output as the gate for marking work complete
Strategy Selection
Choose the smallest strategy that keeps progress recoverable:
| Situation | Strategy |
|---|---|
| Spec-driven planning phase | Write one complete phase artifact, validate it, stop for approval |
| Heavy task decomposition | Decompose into small tasks with dependencies, verification, and traceability |
| Heavy implementation | Treat tasks.md as the durable queue and continue task-by-task |
| Debugging with unclear root cause | Create hypotheses, test them one at a time, record the current best finding |
| Context is getting large | Write a resume checkpoint before continuing |
| Validation fails | Fix if local and clear; otherwise mark blocked with evidence |
Durable Execution Protocol
Use this protocol whenever work could exceed one comfortable response or tool cycle.
- Identify the durable source of truth.
- Break the work into small units that can be verified independently.
- Mark the current unit in progress before editing.
- Complete only that unit or a small batch of tightly related units.
- Run the smallest meaningful verification.
- Persist status immediately after verification.
- Emit a concise progress update when useful.
- Continue until all units are complete, a real blocker appears, or phase rules require approval.
Spec-Driven Usage
Requirements, Design, And Tasks Phases
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 10d ago First seen · 168 lines · 67 tokens per session scan A d608f11fe998
long-running-work-planning is a skill published in the GitHub repository lindoelio/spec-driven-steroids (54 stars, last pushed 1mo ago), licensed MIT. It adds 67 tokens to every session and 1,284 once invoked, about $0.0003 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 skills, from other repositories
artifact-conventions
Defines preservation, format, and section rules for SDD specification artifacts (spec.md, plan.md, tasks.md, checklists). Use when editing feature-artifact files under specs/ / to prevent accidental corruption of cross-referenced IDs, priorities, and gating state.
plan-authoring
Reference material for writing implementation plans (technical context, architecture decisions, data models, API contracts, project-instructions alignment). Loaded on demand by plan-feature; not directly invokable.
task-generation
Reference material with the canonical task-format grammar and decomposition rules for plan-to-tasks expansion. Loaded on demand by generate-tasks; not directly invokable.
adr-authoring
Defines the canonical MADR format, lifecycle rules, numbering policy, and SAD catalog contract for standalone ADRs under specs/adrs/.
implementation-standards
Reference material with coding standards (defensive coding, error handling, testing patterns). Loaded on demand by the Developer sub-agent (.github/agents/developer.md); not directly invokable.
spec-authoring
Reference material for writing product, technical, and operational specifications (work-item priorities, requirement families, success criteria). Loaded on demand by specify-feature; not directly invokable.