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 navraj007in/architecture-cowork-plugin --skill decision-logginggit clone --depth 1 https://github.com/navraj007in/architecture-cowork-pluginWrote 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/navraj007in/architecture-cowork-plugin/decision-logging)<a href="https://agentmods.dev/skills/navraj007in/architecture-cowork-plugin/decision-logging"><img src="https://agentmods.dev/badge/skills/navraj007in/architecture-cowork-plugin/decision-logging/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/navraj007in/architecture-cowork-plugin/decision-logging"><img src="https://agentmods.dev/badge/skills/navraj007in/architecture-cowork-plugin/decision-logging.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 | $0.00022 | $0.02392 |
| Opus 5 | $0.00011 | $0.01196 |
| Sonnet 5 | $0.00004 | $0.00478 |
| Haiku 4.5 | $0.00002 | $0.00239 |
Grade A, and why
decision-logging 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 — 267 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Decision Logging Skill
When commands make architecture decisions (choosing tech stacks, patterns, frameworks, databases, etc.), they MUST log those decisions to _state.json.decisions[] for auditability and future reference.
When to Log a Decision
Log a decision whenever your command:
- Chooses a technology (database, framework, language, messaging system, auth provider)
- Selects an architectural pattern (monolith vs. microservices, clean vs. modular, etc.)
- Determines a trade-off (prioritize cost over latency, scalability over simplicity, etc.)
- Makes a significant strategic choice that affects the system's design
Do NOT log:
- Routine operations (generating files, running tests)
- Configuration of already-decided technologies
- Re-running previous decisions
Examples of decisions to log:
- ✅ "Choose PostgreSQL for primary database" (blueprint command)
- ✅ "Use React over Vue for UI framework" (scaffold command)
- ✅ "Prioritize fast iteration over peak performance" (mvp-scope)
- ✅ "Use monolithic architecture for initial MVP" (architecture design)
- ❌ "Generated auth routes file" (implementation detail, not decision)
- ❌ "Installed dependencies" (mechanical task, not decision)
Decision Structure
Each decision is a JSON object in the decisions[] array:
{
"id": "D-NNN",
"title": "Choose PostgreSQL for primary database",
"rationale": "ACID compliance for order consistency. Team has PostgreSQL expertise. Query flexibility for future analytics features.",
"alternatives_rejected": [
"MongoDB (weak multi-document transactions, overkill for relational data)",
"DynamoDB (cost scales poorly with joins, limited query flexibility)"
],
"trade_offs": "Higher operational complexity than managed cloud DBs. Team must maintain backups. Cold-start slower than serverless options.",
"made_by_command": "blueprint",
"timestamp": "2026-04-24T14:32:10Z"
}
Field Details
| Field | Type | Required | Description |
|---|---|---|---|
id |
string | Yes | Unique decision ID (e.g., "D-001", "D-002"). Increment sequentially within a project. |
title |
string | Yes | Brief decision title (3-10 words). E.g., "PostgreSQL for primary database" |
rationale |
string | Yes | Why this choice? What does it enable? What problem does it solve? (2-5 sentences or bullets) |
alternatives_rejected |
array of strings | Yes | What were the runner-up options and why did they lose? Makes future reconsideration easier. At least 2 alternatives. |
trade_offs |
string | Yes | What are you giving up? What becomes harder or more expensive? (1-3 sentences) |
made_by_command |
string | Yes | Which command made this decision? (e.g., "blueprint", "sdl", "scaffold") |
timestamp |
string | Yes | ISO-8601 timestamp when decision was made (e.g., "2026-04-24T14:32:10Z") |
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 · 267 lines · 22 tokens per session scan A 84d4a760308c
decision-logging is a skill published in the GitHub repository navraj007in/architecture-cowork-plugin (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 22 tokens to every session and 2,392 once invoked, about $0.0001 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
bootstrap-project
Bootstrap a fresh or existing repo with nyann. TRIGGER when the user says "set up this project", "initialize git workflow", "bootstrap this repo", "scaffold this project", "ngyamm this repo", "use my profile" / "apply the nextjs-prototype profile" (profile mode). ALSO trigger on "standard setup" / "usual stack" / "the…
commit
Generate a Conventional Commits message from the staged diff and create the commit after user confirmation. TRIGGER when the user says "commit these changes", "commit this", "generate a commit message", "stage and commit", "commit with CC", "write a commit message", "summarize my diff into a commit", "/nyann:commit".…
iac-apply
Actually APPLY an Infrastructure-as-Code change — the highest-stakes mutator in nyann; it can mutate real cloud infrastructure. Re-runs the plan, shows it, confirms with the user, then invokes bin/iac-apply.sh --apply (adding --confirm-destroy only when the user explicitly confirms a destructive change). UNMISTAKABLY…
retrofit
Audit an existing repo against a profile and offer to fix what's drifted. TRIGGER when the user says "retrofit this repo", "fix this repo's hygiene", "bring this repo into compliance", "remediate drift", "fix what's drifted", "make this repo match the profile", "my repo is half set up, finish it", "I already have some…
setup
First-run onboarding for nyann. Creates the user config directory, collects preferences through AskUserQuestion interactive pickers, and writes /.claude/nyann/preferences.json. TRIGGER when the user says "set up nyann", "configure nyann", "nyann setup", "onboard me", "first time using nyann", "initialize nyann"…
doctor
Run a read-only hygiene + documentation audit on the current repo. TRIGGER when the user says "is this repo healthy", "check hygiene", "audit this repo" (when they mean inspect, not remediate), "what's drifted", "run doctor", "run a health check", "audit docs", "check for doc drift", "what's broken in this project"…