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 LLM-Coding/Semantic-Anchors --skill semantic-contracts-installergit clone --depth 1 https://github.com/LLM-Coding/Semantic-AnchorsWrote 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/llm-coding/semantic-anchors/semantic-contracts-installer)<a href="https://agentmods.dev/skills/llm-coding/semantic-anchors/semantic-contracts-installer"><img src="https://agentmods.dev/badge/skills/llm-coding/semantic-anchors/semantic-contracts-installer/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/llm-coding/semantic-anchors/semantic-contracts-installer"><img src="https://agentmods.dev/badge/skills/llm-coding/semantic-anchors/semantic-contracts-installer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00132 | $0.01404 |
| Opus 5.5 | $0.00053 | $0.00562 |
| Sonnet 5.5 | $0.00026 | $0.00281 |
| Haiku 4.5 | $0.00013 | $0.00140 |
Grade A, and why
semantic-contracts-installer scanned grade A with 1 finding 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 1mo 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- **Runtime:** python3 (used by `install.sh` for settings/JSON manipulation and path resolution); `curl` for `refresh-contracts.sh`. How it starts
The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Semantic Contracts Installer
Install a small, stable set of Semantic Contracts so coding agents start each session with the project's shared vocabulary and rules — not just terms the model already knows, but the compositions and rendered templates it cannot guess.
Why contracts first
A semantic anchor (BLUF, MECE, arc42) already works in any chat — the model knows it; installing a list of anchors adds little. A semantic contract is different: it pins which anchors apply in this project and carries a project-local template (e.g. the "Specification" contract = Cockburn use cases + Gherkin + EARS + activity diagrams, with the full template text). That template is information the model cannot reconstruct on its own. Installing contracts is therefore the primary mode; installing a bare anchor block is the --anchors-only fallback.
This is the Anchor / Contract / Skill taxonomy (ADR-007) in practice: the installer's job is to put the always-on contracts into the agent's instruction file.
Workflow (contracts mode — default)
-
Decide scope and portability. Prefer project-local onboarding for "works across as many agents as possible"; use user-level memory only for agents that support it natively. Read references/agent-support.md before writing files.
-
Present the contract catalog. Read data/contracts.json (a bundled snapshot of the published catalog) and show the user the available contracts by
title+description. Keep the installed set small — usually 2 to 6 contracts that match the project's actual practice. -
Render the selected contracts into a block. For each chosen contract, emit its
template(ortemplateDewhen the project works in German) under a clear heading, followed by its resolvedanchors[]as a one-line supporting-vocabulary list. Start fromassets/templates/contracts-block.md. Do not add the markers — the installer adds them. -
Run the installer instead of editing by hand.
scripts/install.sh --source /path/to/contracts-block.md --target-dir <repo> --scope projectfor repository onboarding.scripts/install.sh --source /path/to/contracts-block.md --scope homefor personal onboarding.- It injects into the first suitable existing markdown file, else creates a minimal
AGENTS.md, and writes an idempotent marker block (<!-- semantic-anchors:start/end -->) so updates replace the previous installation — including any block left by the formersemantic-anchor-onboardingskill.
What ships with it
13 files 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.
- agents/openai.yaml 295 B
- assets/templates/AGENTS.md 689 B
- assets/templates/anchor-block.md 526 B
- assets/templates/CLAUDE.md 546 B
- assets/templates/contracts-block.md 685 B
- assets/templates/copilot-instructions.md 609 B
- assets/templates/GEMINI.md 603 B
- data/contracts.json 47 KB
- references/agent-support.md 4.8 KB
- references/anchor-selection.md 12 KB
- scripts/claude-session-start.sh 1.3 KB runs code
- scripts/install.sh 12 KB runs code
- scripts/refresh-contracts.sh 1.4 KB runs code
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.
- 1mo ago First seen · 75 lines · 132 tokens per session scan A f1b130a63973
semantic-contracts-installer is a skill published in the GitHub repository LLM-Coding/Semantic-Anchors (521 stars, last pushed 3d ago), licensed Apache-2.0. It adds 132 tokens to every session and 1,404 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
distill-feedback
A process for turning corrections you give an agent into lasting working rules. It reviews saved conversations and asks for approval before changing those rules.
lean-ctx
Use LeanCTX Context Gateway tools to select task-relevant code context, apply supported controls, reuse local knowledge and inspect recorded evidence. Use when lean-ctx MCP tools (ctx) are available or file, search or command output needs a focused view.
lean-ctx
Use LeanCTX Context Gateway tools to select task-relevant code context, apply supported controls, reuse local knowledge and inspect recorded evidence. Use when lean-ctx MCP tools (ctx) are available or file, search or command output needs a focused view.
lore
Curating cross-agent knowledge and institutional memory: extracts patterns from agent journals into METAPATTERNS.md, detects knowledge decay, propagates best practices. Use for memory curation.
knowledge-base
Retrieves and updates project-specific prompt knowledge from comparison evidence and user feedback. Use only for prompt analysis or post-comparison learning within a Rashomon evaluation.
add-new-entry-from-temp-md
Format new entries from temp.md into tempentries.md, ready to insert into the section files. USE FOR: adding new resources to the knowledge base. DO NOT USE FOR: editing existing entries or restructuring sections.