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 instructions/dbhq-uk/vela-skill/agents-mdgit clone --depth 1 https://github.com/dbhq-uk/vela-skillWrote 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/instructions/dbhq-uk/vela-skill/agents-md)<a href="https://agentmods.dev/instructions/dbhq-uk/vela-skill/agents-md"><img src="https://agentmods.dev/badge/instructions/dbhq-uk/vela-skill/agents-md.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 | $0.01318 | $0.01318 |
| Opus 5 | $0.00659 | $0.00659 |
| Sonnet 5 | $0.00264 | $0.00264 |
| Haiku 4.5 | $0.00132 | $0.00132 |
Grade A, and why
vela-skill AGENTS.md 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 6d 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 — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI agents (and people) working in this repository.
What this is
vela - compiler-exact code search for .NET, for AI coding agents. It follows the Agent Skills layout (skills/<name>/SKILL.md) and ships as a Claude Code plugin.
Layout
.claude-plugin/plugin.json # plugin manifest
skills/vela/SKILL.md # the skill (agent-facing instructions)
src/Vela/ # the CLI: Config, Harvest, Indexing, Query, Scip
tests/Vela.Tests/ # 426 tests, hermetic
install.sh / install-codex.sh # local installers (Claude / Codex)
docs/ # see docs/README.md for the index
Documentation follows Diataxis. A page is a tutorial
(docs/getting-started.md), a how-to guide (docs/guides/), reference
(docs/reference.md, docs/scip-ecosystem.md) or explanation (docs/architecture.md,
docs/design-notes.md). Mixing modes on one page is the standard failure. Anything new
goes in docs/README.md so every page stays reachable from one place, and
the README stays a shop window rather than a manual.
The three constraints that define this tool
Break any of these and it stops being the thing people can trust:
-
Deterministic only. Every answer follows from Roslyn's semantic model. No model calls, no network, no telemetry, no heuristic ranking. A finding is what the compiler believes, or it is not a finding.
-
Never write to the indexed repository. The index lives outside the source tree. vela reads; it does not modify. Indexing someone's repository must leave it byte-identical.
-
An incomplete index must never look like a complete one. This matters more here than in most tools. If a project fails to load, every query touching it must say so and the exit code must reflect it. An agent that receives an empty reference list will conclude the symbol is unused and delete it. Absence of results is never evidence of absence - report the gap loudly or do not answer.
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.
- 6d ago First seen · 83 lines · 1,318 tokens per session scan A 079a3389d158
vela-skill AGENTS.md is an instructions file published in the GitHub repository dbhq-uk/vela-skill (6 stars, last pushed 2d ago), licensed MIT. It adds 1,318 tokens to every session, about $0.0066 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 2026-08-31.
Other instructions, from other repositories
BlazorBindings.Maui AGENTS.md
AGENTS.md instructions for Dreamescaper/BlazorBindings.Maui, covering mobile guidelines, platform, razor and component conventions and blazor bindings generation.
Connapse CLAUDE.md
Claude Code instructions for Destrayon/Connapse, covering claude.md, build & run, build, run web app (requires backing services) and full stack via docker.
Hosted_Agents_Workshop_dotNET AGENTS.md
Instructions for microsoft/Hosted_Agents_Workshop_dotNET, covering agents.md, repository purpose, general rules, documentation rules and code and build rules.
dotnet-toolkit CLAUDE.md
Instructions for Attemainio/dotnet-toolkit, covering claude.md, what this is, non-negotiable workflow, commands and where to read what.
code-meridian AGENTS.md
Instructions for Driftya/code-meridian, covering agents.md, behavior expectations, required codemeridian usage, trigger rules and repository rules.
code-meridian copilot-instructions.md
Instructions for Driftya/code-meridian: Read AGENTS.md — it is the single source of truth for architecture, conventions, tool usage triggers, and contribution rules for this repository.