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.
git clone --depth 1 https://github.com/manishiitg/coding-agent-loopWrote 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/commands/manishiitg/coding-agent-loop/daily-learnings)<a href="https://agentmods.dev/commands/manishiitg/coding-agent-loop/daily-learnings"><img src="https://agentmods.dev/badge/commands/manishiitg/coding-agent-loop/daily-learnings/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/commands/manishiitg/coding-agent-loop/daily-learnings"><img src="https://agentmods.dev/badge/commands/manishiitg/coding-agent-loop/daily-learnings.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 | $0.00000 | $0.00563 |
| Opus 5.5 | $0.00000 | $0.00225 |
| Sonnet 5.5 | $0.00000 | $0.00113 |
| Haiku 4.5 | $0.00000 | $0.00056 |
Grade A, and why
daily-learnings 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 22d 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Do this Crew's end-of-day learning review: read today's conversations, then save what is new and worth keeping into memory and skills.
{{context}}
1. Read today's conversations
- Today's date is in the context above; if not, use the system date.
- Read every conversation this Crew had today, not just the current chat:
builder/conversation/<YYYY-MM-DD>/— chats with this Crew's owner;builder/crew-chats/users/*/— other users' chats with this Crew (use each file's modification time or message timestamps to find today's).
- Files are JSON:
conversation_history[].Roleand.Parts[].Text. They can be large; skim, search (e.g.jq,grep) and focus on user requests, corrections, decisions, outcomes and repeated procedures. Skip tool output noise. - If a date other than today is given in the context (e.g. "yesterday" or a specific date), review that day instead.
2. Decide what is a learning
Keep only things that are new today and would change how this Crew works next time:
- facts, preferences, decisions, constraints and corrections — test each with "Crew should remember that…";
- procedures that worked (or failed and were fixed) and will recur — test each with "When asked to do X, Crew should…";
- mistakes the user had to correct, with the correct approach.
Leave out transient status, one-off details, raw chat, guesses, unverified research and credentials. If nothing new qualifies, say so and change nothing.
3. Update memory
- Read the project-root MEMORY.md first (create it if missing).
- Follow the project memory contract: concise reverse-chronological entries, one topic per dated heading, merge into an existing topic instead of duplicating, and replace or remove entries that today's evidence contradicts.
4. Update skills
Invoking this command is my explicit ask to preserve reusable procedures:
- Call
list_skillsfirst and read the names and descriptions of existing project skills. - Create or update focused project-local skills at
skills/<skill-name>/SKILL.md(withnameanddescriptionfrontmatter) — never the account-wideskills/custom/library. Update an existing skill only for the same topic and trigger; otherwise create a new, clearly named one; split catch-all skills. - Select each created or updated skill with
update_project_skill_selection.
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.
- 22d ago First seen · 41 lines · 0 tokens per session scan A aeafb67d5cc6
daily-learnings is a command published in the GitHub repository manishiitg/coding-agent-loop (13 stars, last pushed 19d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 563 tokens. 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 2026-09-25.
Other commands, from other repositories
blackboard
Read or write a Network-AI blackboard key (shared multi-agent state).
go
You MUST start your FIRST response with this exact agent activation line.
dashboard-session-fork
Fork a session into a new one. Usage /dashboard:session-fork.
dashboard-session-hide
Hide a session from the dashboard list. Usage /dashboard:session-hide.
resolve-dependabot-prs
Triage open Dependabot PRs, return a quick risk/action summary table, then enter planning mode with an ordered rebase→merge→follow-up→resume resolution plan.
relay
Create or update a root relay.md session-continuation document.