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 Osipchuk/agent-skills --skill adr-capturegit clone --depth 1 https://github.com/Osipchuk/agent-skillsWrote 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/osipchuk/agent-skills/adr-capture)<a href="https://agentmods.dev/skills/osipchuk/agent-skills/adr-capture"><img src="https://agentmods.dev/badge/skills/osipchuk/agent-skills/adr-capture/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/osipchuk/agent-skills/adr-capture"><img src="https://agentmods.dev/badge/skills/osipchuk/agent-skills/adr-capture.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00181 | $0.00770 |
| Opus 5 | $0.00090 | $0.00385 |
| Sonnet 5 | $0.00036 | $0.00154 |
| Haiku 4.5 | $0.00018 | $0.00077 |
Grade A, and why
adr-capture 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 9d 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
adr-capture
Decisions get lost because writing them down is friction. This skill catches the decision while it's warm and produces an ~80%-complete draft for review. It never auto-commits, and it biases toward false negatives — better to miss a decision and be asked than to spam the ADR log with drafts of things that were only floated.
Detecting a decision
A real decision has two parts: a commitment signal and a rationale.
- Commitment: "going with", "we'll use", "decided", "chose X over Y", "let's standardize on".
- Rationale: at least one because/why, even if implicit.
Exploratory phrasing — "might", "could", "what if", "worth considering" — is NOT a decision. If you see commitment without a clear rationale, or you're unsure, ask one question to confirm before drafting. Do not draft on ambiguity.
Drafting workflow
-
Config. Read optional
.claude/adr/config.yamlforadr_dir(defaultdocs/adr/), the id format, and the list ofmandatory_sections— seereferences/config.example.yamlfor the schema. Otherwise use the defaults baked intoreferences/adr-template.md. -
Next id. Scan
adr_dirfor the highest existing ADR number and increment. Slugify the decision title for the filename, e.g.0042-vector-store.md. -
Fill from the conversation. Map what was said onto the template: context and forces, the decision, alternatives considered and why each was rejected, consequences (both + and −).
-
Mark the gaps. For every mandatory section the conversation did not cover, insert a literal
<TODO: ...>with a hint of what's needed. Never invent regulatory, security, data-classification, or rollback content — those must come from a human. -
Review, then write. Show the full draft in chat first. Write the file only after explicit confirmation, with
Status: Proposed. Never commit, and never set any status past Proposed on the user's behalf — that is a human approval step.
What ships with it
2 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.
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.
- 9d ago First seen · 57 lines · 181 tokens per session scan A 83ab73aff54c
adr-capture is a skill published in the GitHub repository Osipchuk/agent-skills (5 stars, last pushed 29d ago), licensed MIT. It adds 181 tokens to every session and 770 once invoked, about $0.0009 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 skills, from other repositories
opensrc
Fetch dependency source code to give AI agents deeper implementation context. Use when the agent needs to understand how a library works internally, read source code for a package, fetch implementation details for a dependency, or explore how an npm/PyPI/crates.io package is built. Triggers include "fetch source for"…
lx
Codebase exploration tool that reads many files or whole directories in a single call, with per-file headers, glob include/exclude filters, function/type skeleton extraction (signatures only, no bodies), and head/tail line slicing.
audit
Audit phase. Parallel review: code quality + security + tests. Semantic dedup of cross-mode findings. Outputs PASS/WARN/FAIL per dimension. Validates spec coverage.
discover
Trigger: vague/unfocused request or solution-without-problem. Also invoked via /discover command. Reframes goal before acting.
orbit
State-persisted autonomous pipeline: spec → go → audit → eval → ship → evolve in one command. Auto-detects direct/council/interactive mode. Crash-recoverable via PIPELINE-.json. Hands-off until PR.
reflect
On-demand human self-assessment of AI usage quality. Scores 5 dimensions from session data. Not agent performance review.