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 sh5623/self-improvement --skill si-initgit clone --depth 1 https://github.com/sh5623/self-improvementWrote 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/sh5623/self-improvement/si-init)<a href="https://agentmods.dev/skills/sh5623/self-improvement/si-init"><img src="https://agentmods.dev/badge/skills/sh5623/self-improvement/si-init/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/sh5623/self-improvement/si-init"><img src="https://agentmods.dev/badge/skills/sh5623/self-improvement/si-init.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 | $0.00040 | $0.00654 |
| Opus 5 | $0.00020 | $0.00327 |
| Sonnet 5 | $0.00008 | $0.00131 |
| Haiku 4.5 | $0.00004 | $0.00065 |
Grade A, and why
si-init 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.
What it actually says
Initialize self-improvement in Codex
Read runtime boundaries and data discovery and format maps.
Identify the project root and the canonical existing system before creating a file. Existing Claude registrations and table-less convention logs remain canonical. Do not copy a whole CLAUDE.md into AGENTS.md, create a Claude import, or start a second log for Codex.
The bundled helper performs conservative, idempotent wiring:
python3 "<this-skill-directory>/../../scripts/init_project.py" --project "<project-root>" --check
python3 "<this-skill-directory>/../../scripts/init_project.py" --project "<project-root>"
Resolve the script relative to this installed SKILL.md, not to the shell cwd. Python 3.9+ is required only for the helper. Without it, perform the same inspection and minimal edits directly; report that the helper was not run.
- The helper creates a neutral generated log only when no credible system exists. It
registers existing data without modifying its bytes. Multiple candidates require an
explicit canonical choice:
--data-file "docs/actual-log.md". - Existing registered systems default to
archive: noneuntil their actual destination is confirmed. Preserve a legacy declared archive or pass--archive "docs/actual-archive"; this registers the path and does not move records. - It writes only the marked Codex section in the effective AGENTS.override.md/AGENTS.md. Symlinked instruction files require inspecting their ownership first; do not bypass a refusal by writing through a link into CLAUDE.md.
- Inspect the reported format map and correct it to the real system's sections, log unit,
budgets and archive.
noneis valid. Do not create missing tables just to satisfy defaults.
Inspect the actual document terrain: tool configs and scripts; nested Codex instruction
files; shared task/domain docs; Claude-only files as a separate runtime layer. On a new log,
fill §routing with existing locations or none (create when needed). Add Codex destinations
to an existing routing system in its own format, without changing Claude's routing rows.
After a plugin update, compare the installed .codex-plugin/plugin.json version with the Codex template stamp, if present. The helper intentionally preserves existing data. Inspect stale wording and repair only Codex-owned fixed text when needed, preserving table values, index, migration history, log bodies and all Claude stamps. A public/internal version switch is not evidence that the data needs upgrading; inspect schema and content.
Report created / registered / unchanged / repaired, the real data path, effective Codex
instruction path, format map and any unmeasured or missing layer. Confirm Claude files
were not modified by Codex wiring. Next action: $self-improvement:si-improve when an evidenced gap recurs.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 Changed · -103 lines · -75 tokens per session scan C → A 081234b2a55d
- 6d ago Changed · -2 tokens per session 93a51982c1a2
- 10d ago First seen · 155 lines · 117 tokens per session scan C e7086137388d
si-init is a skill published in the GitHub repository sh5623/self-improvement (1 stars, last pushed 4d ago), licensed MIT. It adds 40 tokens to every session and 654 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 2026-08-31.
Other skills, from other repositories
recursive-spine-pollinate
Use when something just proved itself in real work and should be carried to other projects — captures it as a pollen record in the builder's hive (registry file + paired issue), or, in pull mode, reads every configured hive and offers relevant transplants into the current repo, recording each transplant back onto the…
ucai-patterns
Use when the user asks about Claude Code best practices, how to write agents, how to use hooks, how to manage context, or how to work effectively with Claude Code's native systems.
memory-curator
Use at session-close when a session captured several new memories, or on-demand, to consolidate, deduplicate, prune, and structure the beads memory store. Triggers on "curate memories", "clean up memories", "memory sweep".
orient
Use when user invokes /orient with a topic keyword, entity type, project name, time qualifier, or combination. Also triggers on "what do we know about X", "remind me about X", "where did we leave off on X". Provides targeted context loading — searches the MCP Memory Server graph, knowledge files, journal entries, and…
memory
Use when user invokes /memory with a subcommand (search, trace, add, episode, status). Also triggers on "search my memory for X", "add to memory", "what do I know about X" (when membrain is available), "memory status", "memory trace". Searches and manages the membrain knowledge graph via MCP tools or HTTP API.
capture
Use when user invokes /capture to capture information from the current conversation into a staging file for later reflection and knowledge graph processing. Accepts no arguments (auto-extract), specific facts as text, or filter keywords like "decisions" or "technical". Do NOT use when the user wants to write directly…