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 white1024/agent-nextup --skill retrospectgit clone --depth 1 https://github.com/white1024/agent-nextupWrote 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/white1024/agent-nextup/retrospect)<a href="https://agentmods.dev/skills/white1024/agent-nextup/retrospect"><img src="https://agentmods.dev/badge/skills/white1024/agent-nextup/retrospect/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/white1024/agent-nextup/retrospect"><img src="https://agentmods.dev/badge/skills/white1024/agent-nextup/retrospect.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 | $0.00069 | $0.01111 |
| Opus 5.5 | $0.00028 | $0.00444 |
| Sonnet 5.5 | $0.00014 | $0.00222 |
| Haiku 4.5 | $0.00007 | $0.00111 |
Grade A, and why
retrospect 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 4d 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 — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
retrospect — improve the ground the next session stands on
Goal: every friction this stretch of work hit lands in exactly one destination that changes what the next session does — and the mechanical ones land as checks, not sentences.
1. Gather the evidence
- Call
post_mortem_candidatesfirst. It derives candidates from hard ledger signals — overrides, blocked tasks, failed tool calls, refused calls — with the evidence attached. These are the floor: every one of them gets a verdict. - Then read the rest of the stretch: recent ledger entries, tasks that were reopened or blocked, superseded decisions, and the conversation if you are still in it.
- Read what agents here actually load: AGENTS.md outside the engine's block, the module guides in
nextup_docs/modules/, memory, andlessonsin.nextup/rules.json. - Answer the project's own retrospective questions as you go —
flywheel.postMortemPromptsin.nextup/rules.json.
Done when every candidate from the first call carries a note of what actually happened, and every prompt has an answer.
2. Classify each finding
Ask these in order; the first yes decides the destination.
- Could a machine have caught it? A wrong command, a broken link, a forbidden pattern, a value out of range, a file in the wrong place. The fix is a check — a test, a lint rule, a hook, a CI step. Look for the project's own check commands first: one that exists but is not wired in, or passes without checking anything, is itself the finding. A project with no automated check at all is a finding too.
- Was the information out of reach? A log nobody could see, a server's output, a document the agent never found. The fix is access, or a pointer: one line in the right steering file naming where to look.
- Was it a judgement call that went the wrong way? A lesson — an actionable rule for next time.
- Was a line crossed that must never be crossed? A boundary.
- Was an approach tried and rightly dropped? A rejection, so nobody proposes it again.
- Was a tool call slow or wasteful? A lesson or a pointer naming the cheaper route.
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.
- 4d ago First seen · 74 lines · 69 tokens per session scan A 7d9a2775b34d
retrospect is a skill published in the GitHub repository white1024/agent-nextup (2 stars, last pushed 5d ago), licensed MIT. It adds 69 tokens to every session and 1,111 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-10-19.
Other skills, from other repositories
lossless-openclaw-orchestrator
Use LCO from an OpenClaw agent to search, describe, expand, and dry-run local Codex session work without reading raw transcripts.
Google Workspace Expert
Deep Google Sheets, Gmail, and Calendar knowledge for business operators — formulas, search operators, automation patterns, and cross-tool workflows. Not a setup guide — domain expertise for getting real work done.
Notion Expert
Deep Notion knowledge for business operators — databases, templates, property design, formulas, and automation patterns. Not a setup guide — domain expertise for getting real work done in Notion.
pack-codex-move
A workflow for collecting and packaging a user’s local Codex setup before moving to another computer. Codex is an AI coding assistant, and the package is prepared for later restoration.
n8n-node-configuration
Operation-aware node configuration guidance. Use when configuring nodes, understanding property dependencies, determining required fields, choosing between getnode detail levels, or learning common configuration patterns by node type.
aident-loadout
Run multi-step work across apps the user connected in Aident Loadout: live action schemas, connection checks, cost preflight, and approval before side effects.