self-journal

self-journal is an agent for coding agents from cpuguy96/StepCOVNet. It costs 0 tokens per session (11,220 once invoked), scanned A, original, Apache-2.0.

A hand-maintained journal of improvements to how an agent works on the project. It records process lessons separately from research results and experiment records.

In plain words
What is it for?
Recording steering corrections, documenting process fixes, and linking each lesson to a rule, skill, or code change made during the same work session.
Why use it?
It preserves useful corrections and workflow decisions so the same mistakes or delays are less likely to happen again.

Agent

Install

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.

agentmods
npx agentmods add agents/cpuguy96/stepcovnet/self-journal
Clone the repo
git clone --depth 1 https://github.com/cpuguy96/StepCOVNet

Wrote 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.

agentmods badge for self-journal

README.md
[![agentmods](https://agentmods.dev/badge/agents/cpuguy96/stepcovnet/self-journal.svg)](https://agentmods.dev/agents/cpuguy96/stepcovnet/self-journal)
Your own site
<a href="https://agentmods.dev/agents/cpuguy96/stepcovnet/self-journal"><img src="https://agentmods.dev/badge/agents/cpuguy96/stepcovnet/self-journal.svg" alt="Measured on agentmods" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 11,220 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.11220
Opus 5 $0.00000 $0.05610
Sonnet 5 $0.00000 $0.02244
Haiku 4.5 $0.00000 $0.01122

Measured 4d ago against content hash d76c393a177d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

self-journal 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.

docs/agents/self-journal.md · 578 lines

How it starts

The opening of the file, as written. The whole thing — 578 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent self-journal

Iterative improvements to agent behavior, process, and conventions on this project. Not a substitute for EXPERIMENT_LOG.md or DISCUSSION_NOTES.md — those hold research findings; this holds how we work better.

Order: newest entries first (prepend below).

Same-turn promotion: The journal is a receipt, not the fix. Every entry must cite an Artifact path (rule, skill, or code) created or updated in the same turn before the JRN is written. See agent-self-improvement skill.

When to prepend

Immediately when the user makes a steering correction — how you decide, prioritize, run commands, log output, route docs, commit, etc. Do not wait for a long or “substantive” session to end.

Also when you discover:

  • A mistake that wasted time or produced wrong conclusions
  • A fix or workaround that should become default
  • A user-established convention worth preserving
  • A repeated workflow missing from .cursor/skills/

Periodically during work: after discrete tasks with process lessons, when the same friction recurs, or before switching task areas — journal in the same session; do not defer everything to session end.

Maintenance rule: artifact first, journal second in the same turn; link EXP-… / NOTE-… when relevant.

Entry format

Insert at the top of Entries (below this section):

### JRN-YYYYMMDD-NN: Short title

| Field            | Value                                            |
| ---------------- | ------------------------------------------------ |
| **Timestamp**    | YYYY-MM-DD HH:MM:SS (system clock at write time) |
| **Category**     | mistake \| fix \| convention \| skill-gap        |
| **Summary**      | What happened                                    |
| **Artifact**     | Path(s) created or updated (required)            |
| **Action taken** | One line: what the artifact enforces             |
| **Related**      | EXP-…, NOTE-…, skill name                        |

Read the full file on GitHub · 578 lines

Changes

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.

  1. 4d ago First seen · 578 lines · 0 tokens per session scan A d76c393a177d

Subscribe to this mod's changes

self-journal is an agent published in the GitHub repository cpuguy96/StepCOVNet (22 stars, last pushed 10d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 11,220 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 2026-08-30.