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 skills/outlinedriven/outline-driven-development/abstract-wellnpx skills add OutlineDriven/outline-driven-development --skill abstract-wellgit clone --depth 1 https://github.com/OutlineDriven/outline-driven-developmentWrote 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/outlinedriven/outline-driven-development/abstract-well)<a href="https://agentmods.dev/skills/outlinedriven/outline-driven-development/abstract-well"><img src="https://agentmods.dev/badge/skills/outlinedriven/outline-driven-development/abstract-well.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 5.1 | $0.00059 | $0.00899 |
| Opus 5 | $0.00030 | $0.00449 |
| Sonnet 5 | $0.00012 | $0.00180 |
| Haiku 4.5 | $0.00006 | $0.00090 |
Grade A, and why
abstract-well 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- abstract-well — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Abstract well
Contract
| Field | Bound contract |
|---|---|
| Trigger | The user wants to classify a set of abstractions as useful, bad, or busy and keep one shallow level. |
| Authority | Read-only. No file, VCS, credential, paid, published, deployed, or remote mutation. |
| Side effect | A chat-output classification and one-shallow-level recommendation. Nothing is written or changed. |
| Done | Every abstraction in scope is classified as useful, bad, busy, or unclassified, and a single shallow level is selected whose cooked form is easier than its raw form, or an explicit statement is returned that no useful level exists. |
Inputs
- The abstractions to review. Supply either an explicit list of named abstractions (wrappers, helpers, interfaces, layers, adapters) or a scope (file, module, or function set) from which they are enumerated.
- The caller perspective: whose code uses each abstraction. If omitted, the caller is the immediate consumer of the abstraction.
- The underlying operation each abstraction wraps must be knowable so a raw form can be written. If it is not, the abstraction is marked unclassified.
Procedure
- Enumerate the abstractions in scope. If the user named them, use that list. If the user gave a scope, list every named wrapper, helper, interface, layer, or adapter inside it. Do not invent abstractions that the input does not name or contain. Done when: every abstraction in scope is listed.
- For each abstraction, write two concrete forms at the same caller task: the raw form, the code the caller writes without the abstraction, and the cooked form, the code the caller writes with it. Keep both at the same task and the same caller. Done when: raw and cooked forms exist for every abstraction.
- Classify each abstraction by comparing cooked form to raw form for that caller:
- useful: the cooked form is easier than the raw form — fewer concepts, less ceremony, fewer places to get wrong.
- bad: the cooked form is harder than the raw form — more indirection, more concepts, or more ceremony than the inline path.
- busy: the cooked form is neither easier nor harder — it relocates complexity to another place without reducing total difficulty for the caller.
- unclassified: the raw form cannot be written concretely because the underlying operation is unknown. Done when: every abstraction is classified as useful, bad, busy, or unclassified.
- Select one shallow level to keep: the single useful abstraction whose cooked-form advantage over its raw form is largest. If no abstraction is useful, select none and state that explicitly. Done when: one useful abstraction is selected or an explicit statement that none exists.
- Reject stacking: when two abstractions both sit between the same caller and the same raw operation, keep only the one with the larger cooked-form advantage; do not recommend keeping both layers. Done when: no two stacked abstractions both remain between the same caller and raw operation.
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.
- 2d ago First seen · 44 lines · 59 tokens per session scan A f006b0364d4a
abstract-well is a skill published in the GitHub repository OutlineDriven/outline-driven-development (52 stars, last pushed yesterday), licensed Apache-2.0. It adds 59 tokens to every session and 899 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 2026-09-03.
Other skills, from other repositories
outlines
Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local models (Transformers, vLLM), and maximize inference speed with Outlines - dottxt.ai's structured generation library.
brainstorm
Explore vague or ambitious ideas into a right-sized requirements-only plan. Use when the user wants to brainstorm, think through scope, decide what to build, or needs collaborative product framing before planning, not for a decisive verdict on whether to adopt or switch to a specific external technology, library, or…
doc-review
Use when the user asks to review or critique a prose planning document — a plan, spec, PRD, requirements doc, or design doc.
drift-detect
Use when the user says "plan drift", asks whether the roadmap, plans, or docs still match the code, or is deciding what to rebuild when restarting a stalled project. For doc-vs-code drift inside a specific diff, use sync-docs.
enhance
Audit and tighten agent/plugin surfaces. Use when the user asks to enhance a plugin config, agent definition, skill, or CLAUDE.md.
git-branchless
Use when asked for multi-commit stack edits, rebases, fixups, or stacked-PR publishing with branchless git idioms and a publish gate. Not for plain-git workflows in repos without branchless; the skill stays inert there. Human-only.