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 avelikiy/great_cto --skill decision-evalgit clone --depth 1 https://github.com/avelikiy/great_ctoWrote 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/avelikiy/great_cto/decision-eval)<a href="https://agentmods.dev/skills/avelikiy/great_cto/decision-eval"><img src="https://agentmods.dev/badge/skills/avelikiy/great_cto/decision-eval/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/avelikiy/great_cto/decision-eval"><img src="https://agentmods.dev/badge/skills/avelikiy/great_cto/decision-eval.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 | $0.00039 | $0.00916 |
| Opus 5 | $0.00019 | $0.00458 |
| Sonnet 5 | $0.00008 | $0.00183 |
| Haiku 4.5 | $0.00004 | $0.00092 |
Grade A, and why
decision-eval scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Decision Eval — automated scoring for architectural alternatives
Invoke after architect proposes 2+ variants, before creating gate:arch.
When to invoke
Invoke this skill when ALL of these are true:
- An ADR (
docs/adr/ADR-*.md) or ARCH doc (docs/architecture/ARCH-*.md) contains a section with 2 or more named alternatives (look for## Alternatives Considered,## Options, or bold-prefixed options like**Option A:**) - The architect has not yet created
gate:arch - The user has not said "skip scoring", "no scoring", or "skip decision-eval"
project_sizein PROJECT.md is NOTnano
Skip silently (do not even mention) if any condition fails.
How to invoke
Read the most recent ADR or ARCH doc to confirm 2+ variants exist, then spawn
the decision-scorer agent with the file path as context:
# Identify target document
TARGET=$(ls -t docs/adr/ADR-*.md 2>/dev/null | head -1)
[ -z "$TARGET" ] && TARGET=$(ls -t docs/architecture/ARCH-*.md 2>/dev/null | head -1)
# Confirm 2+ variants
VARIANT_COUNT=$(grep -cE "^\*\*[A-Za-z]|^### [A-Za-z]|^- \*\*[A-Za-z]" "$TARGET" 2>/dev/null || echo 0)
If VARIANT_COUNT >= 2, dispatch the agent:
Agent: decision-scorer
Context: <TARGET file path>
Task: Score the architectural variants in <TARGET> against .great_cto/PROJECT.md criteria.
Save output to docs/decisions/.
Output location
The decision-scorer agent saves results to:
docs/decisions/DECISION-<slug>-<YYYYMMDD>.md
After the agent completes, read the output file and surface the recommendation to the architect:
Decision scoring complete:
Recommended: <variant name> (<score>/5.00)
Runner-up: <variant name> (<score>/5.00)
Full report: docs/decisions/DECISION-<slug>-<YYYYMMDD>.md
Architect: review the scoring rationale before accepting or overriding the recommendation.
Skip conditions
Output nothing and proceed to the next step if:
project_size: nanoin PROJECT.md- Fewer than 2 variants found in the ADR/ARCH doc
- User message contains "skip scoring" or "skip decision-eval" or "no scoring"
- The target document is a bug-fix or docs-only ADR (check title for "fix:", "docs:", "chore:")
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.
- 7d ago First seen · 103 lines · 39 tokens per session scan A bd3053e02e25
decision-eval is a skill published in the GitHub repository avelikiy/great_cto (92 stars, last pushed today), licensed MIT. It adds 39 tokens to every session and 916 once invoked, about $0.0002 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
spec-drift
Standalone plan-vs-code audit on any branch: runs /ship Step 8's plan-completion section from disk (hash-pinned) against an explicit plan and base. Report, JSON, exit code. Never edits code.
third-lens-review
After Claude self-pitfall + Codex on a ship-worthy/architecture/RT/security/contract change: run a third external model house (distant training distribution → different blind spots) on the patched artifact, then adversarial synthesis.
quality-review
After a PRD, spec, or plan, before implementation: hunt perceived-quality pitfalls (silent failures, missing loading/empty states, error recovery, state drift) that make a product feel cheap. Complements pitfall-verification.
boundaries
Analyze Phoenix context boundaries and module coupling via mix xref. Use when checking cross-context calls, validating dependencies, before splitting modules, or reviewing architecture.
triage
Triage review findings interactively — approve, skip, or prioritize each issue. Use after /phx:review to filter findings before fixing.
review-code
Review a change along one specific quality dimension — bugs, design, simplicity, maintainability, testability, test quality, type safety, contracts, operational readiness, docs, prose value, change intent, defect-class completeness after a fix, or context-file adherence. Loads exactly one dimension reference and…