Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add alexmond/alexmskills/plugin install evolving-claude-mdWrote 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/alexmond/alexmskills/evolving-claude-md)<a href="https://agentmods.dev/skills/alexmond/alexmskills/evolving-claude-md"><img src="https://agentmods.dev/badge/skills/alexmond/alexmskills/evolving-claude-md/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/alexmond/alexmskills/evolving-claude-md"><img src="https://agentmods.dev/badge/skills/alexmond/alexmskills/evolving-claude-md.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 | $0.05008 |
| Opus 5 | $0.00000 | $0.02504 |
| Sonnet 5 | $0.00000 | $0.01002 |
| Haiku 4.5 | $0.00000 | $0.00501 |
Grade A, and why
evolving-claude-md 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 9d 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 — 331 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Evolving CLAUDE.md
Try it:
/evolving-claude-md:evolving-claude-md— or say "make CLAUDE.md evolve".
CLAUDE.md is read into Claude's context on every turn. Every byte costs tokens. So the goal isn't "log everything we learned" — it's a small, well-pruned set of durable decisions that future-Claude needs, with everything else linked to or archived.
This skill wires that. Three mechanisms keep it healthy automatically:
| Hook | When | What it does |
|---|---|---|
| SessionStart | start of every session | audit-claude-md.py — if the D&L section is bloated, injects a recommendation to compact |
PreToolUse on Write|Edit of CLAUDE.md |
before each edit | lint-claude-md.py — rejects new entries that violate the format (no topic tag, no date, > 200 chars) |
| PostCompact | after context compaction | re-runs the audit so the assistant sees the current D&L state without paying for the full file twice |
When installed as a plugin, the three hooks ship inside the plugin (hooks/hooks.json, pathed via ${CLAUDE_PLUGIN_ROOT}) and register automatically once enabled — nothing to add to your settings. For a manual install, copy the three scripts to .claude/skills/evolving-claude-md/ and add the hooks to .claude/settings.json (see Setup checklist). Disable individually by removing the entry; disable all via disableAllHooks: true in settings.
Not for documenting a codebase from scratch — that's
init. This adds the mechanism that keeps the file current and bounded once it exists.
Format — the contract every entry must follow
- YYYY-MM-DD — **topic-tag** — short statement. Why: brief reason. [Optional: see → docs/decisions/...].
- YYYY-MM-DD — calendar date, no relative dates.
**topic-tag**— kebab-case, one or two words, MANDATORY. Reuse existing tags where they fit; the lint hook surfaces the inventory. Pick a stable vocabulary per project (e.g.auth,build,schema,ci,perf).- One sentence of what. The why is the load-bearing half — lead with constraint, incident, or preference.
- Hard cap: 200 chars in the body, max 3 lines. Bigger? Move the detail to
docs/decisions/{YYYY-MM-DD}-{topic}.mdand keep the entry as a one-line teaser linking there. The lint hook enforces this.
What ships with it
7 files 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.
- 9d ago First seen · 331 lines · 0 tokens per session scan A 3639d07cfe6f
evolving-claude-md is a skill published in the GitHub repository alexmond/alexmskills (6 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 5,008 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 2026-08-31.
Other skills, from other repositories
continuous-learning-v2
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.
context-engineering
Manage, optimize, and debug Claude Code context windows. Use this skill whenever the user mentions context limits, context window, autocompaction, compaction firing, context rot, losing context mid-session, session degradation, running out of tokens, /compact, CLAUDEAUTOCOMPACTPCTOVERRIDE, context compression…
memory-persistence
Implement cross-session memory, persistent agent knowledge, and session continuity for Claude Code. Use this skill whenever the user wants Claude to remember things between sessions, asks how to avoid re-explaining context each day, wants to build a memory system for an agent, mentions CLAUDE.md memory, session files…
hive.note-taking
Maintain a free-form scratchpad of decisions, extracted values, and open questions so context pruning doesn't lose anything you still need.
security-compliance
Guides security professionals in implementing defense-in-depth security architectures, achieving compliance with industry frameworks (SOC2, ISO27001, GDPR, HIPAA), conducting threat modeling and risk assessments, managing security operations and incident response, and embedding security throughout the SDLC.
stride-analysis-patterns
Apply STRIDE methodology to systematically identify threats. Use when analyzing system security, conducting threat modeling sessions, or creating security documentation.