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 tmj-90/gaffer --skill slo-architectgit clone --depth 1 https://github.com/tmj-90/gafferWrote 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/tmj-90/gaffer/slo-architect)<a href="https://agentmods.dev/skills/tmj-90/gaffer/slo-architect"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/slo-architect/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/tmj-90/gaffer/slo-architect"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/slo-architect.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 | $0.00084 | $0.01013 |
| Opus 5 | $0.00042 | $0.00507 |
| Sonnet 5 | $0.00017 | $0.00203 |
| Haiku 4.5 | $0.00008 | $0.00101 |
Grade A, and why
slo-architect 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Define SLOs that mean something
Most "SLOs" in the wild are arbitrary numbers no one believes — 99.9% on every endpoint, no SLI definition, no error budget, no policy for when budget burns. This skill enforces the discipline from Google's SRE Workbook.
Four cardinal mistakes
- Target too high (99.99%+ on services that can't support it) — every minor blip violates; alerts become noise.
- Wrong SLI (CPU usage as proxy for user experience) — system green while users suffer.
- No error-budget policy — burning budget means nothing if there is no agreed action.
- Single-window burn-rate alert — either too noisy (page on a 5-min spike) or too slow (notice budget exhausted after the fact).
Core vocabulary
SLI → measurable signal of user-perceived health (e.g. HTTP 2xx rate, p99 latency)
SLO → target for the SLI over a rolling window (e.g. 99.9% over 30 days)
EB → error budget: (100% − SLO%) × window = how much "bad" you can spend
BR → burn rate: how fast you're consuming the error budget right now
Steps
- Pick the right SLI. Choose a measurement that reflects user experience, not system internals. Event-based (good events / total events) is usually cleaner than time-window averages.
- Set a believable target. Measure your actual reliability first. Set the SLO at or below the 10th percentile of your measured per-window reliability (a level you already meet in ~90% of windows) so it's meaningful but achievable. 99.9% on a service that regularly drops to 99.5% is theatre.
- Calculate the error budget. For 99.9% over 30 days: budget = 0.1% × 30d = 43.2 minutes of downtime. Document this number explicitly.
- Wire multi-window burn-rate alerts. Two windows (short + long) with two burn rates. Canonical Google SRE thresholds: 2% budget in 1h (fast burn, page now) + 5% budget in 6h (slow burn, ticket). Adapt to your SLO window.
- Write the error-budget policy. What happens when >50% of budget is gone mid-window? Freeze feature work, hold risky deploys, escalate. Get agreement before the SLO ships.
- Set a review cadence. Review SLOs quarterly: are they still meaningful? Are they achievable? Do they map to what users actually care about?
- Verify + evidence. Run burn-rate alert thresholds against a replay of the last incident; confirm the fast-burn alert would have fired within 5 min of the outage start. Record output via
record-evidence.
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 · 62 lines · 84 tokens per session scan A af5f0e33e2e1
slo-architect is a skill published in the GitHub repository tmj-90/gaffer (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 84 tokens to every session and 1,013 once invoked, about $0.0004 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
bulwark-brainstorm
Role-based brainstorming with dual modes: --scoped (sequential Task tool, 5 roles) and --exploratory (Agent Teams peer debate, 4 roles). Use for feasibility assessment and idea validation.
plan-creation
Create structured implementation plans via a 4-role scrum team (Product Owner, Architect, Eng/Delivery Lead, QA/Critic) with optional Agent Teams peer debate mode.
anthropic-validator
Validates Claude Code assets (skills, hooks, agents, commands, MCP servers, plugins) against official Anthropic standards. Fetches latest docs dynamically and produces structured validation reports.
create-subagent
Generates single-purpose Claude Code sub-agents for use via the Task tool. Use when creating dedicated sub-agents, scaffolding agent definitions, or generating agents with diagnostics and permissions setup.
test-audit
Audit test suites for T1-T4 violations using AST analysis, mock detection, and multi-stage synthesis. Invoke when user asks to audit tests, check test quality, find mock violations, review test effectiveness, or inspect test suites for over-mocking. Triggers automatic rewrites when quality gates fail.
code-review
Comprehensive code review with distinct aspect based sections. Use when reviewing code, checking for security issues, finding type safety problems, auditing code quality, or when user asks to review code, PRs or changes. Three-phase workflow runs static tools, LLM judgment, and writes diagnostic log.