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 chan4lk/specclaw --skill learngit clone --depth 1 https://github.com/chan4lk/specclawWrote 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/chan4lk/specclaw/learn)<a href="https://agentmods.dev/skills/chan4lk/specclaw/learn"><img src="https://agentmods.dev/badge/skills/chan4lk/specclaw/learn/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/chan4lk/specclaw/learn"><img src="https://agentmods.dev/badge/skills/chan4lk/specclaw/learn.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00017 | $0.00423 |
| Opus 5.5 | $0.00007 | $0.00169 |
| Sonnet 5.5 | $0.00003 | $0.00085 |
| Haiku 4.5 | $0.00002 | $0.00042 |
Grade A, and why
learn 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 15d 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
specclaw learn
First, run specclaw-ensure-init .specclaw — idempotently creates .specclaw/ if it doesn't exist (silent if already initialized; auto-inits using the current directory's basename as the project name).
Capture build learnings — spec gaps, design misses, and patterns discovered during implementation.
Log a learning:
specclaw-log-learning .specclaw <change> <category> <priority> "<detail>" ["<action>"]
Categories: spec_gap | design_gap | pattern | best_practice | agent_issue
Priorities: low | medium | high
List learnings for a change:
specclaw-log-learning .specclaw <change> --list
Promote a learning to the repo-local knowledge base:
specclaw-log-learning .specclaw <change> --promote <id>
Promote writes to .specclaw/knowledge/ — never to the plugin itself:
spec_gap,design_gap→.specclaw/knowledge/spec-guidelines.mdpattern,best_practice,agent_issue→.specclaw/knowledge/agent-hints.md
The build agent receives agent-hints.md as context automatically on every task.
When to log:
- A build revealed a spec gap (requirements unclear or missing)
- A design decision needed mid-build adjustment
- Agents discovered a useful pattern worth reusing
- Parallel tasks created conflicts (duplicate code, shared deps)
- An agent struggled with the context or instructions
Learnings are stored in .specclaw/changes/<change>/learnings.md and feed the pattern detection system for cross-change analysis (see /specclaw-extras:patterns).
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.
- 15d ago Changed · -38 tokens per session b2638db35b3e
- 1mo ago First seen · 43 lines · 55 tokens per session scan A b22713ae61a6
learn is a skill published in the GitHub repository chan4lk/specclaw (13 stars, last pushed 8d ago), licensed MIT. It adds 17 tokens to every session and 423 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
kai
Kai — Self-Improving Meta-Agent. Reviews working-rules a human has already approved in whatever memory backend the project uses, keeps the ones that are universal rather than project-specific, and emits proposed SKILL.md edits as a diff for review. Never applies or commits a change. Use when reviewing accumulated…
debrief
Skill "debrief" from adamlinscott/claude-skills, covering debrief (beta), 1. refresh the corpus, debrief corpus --global # or roll up across all projects, 2. consolidate (grouping-task -> merge / group) and 3. pick a theme and ask the open-question set.
brief-me
The user has come back to this session with weak or no memory of it. Reading the last message will not fix that: the last message was written for someone who was here five minutes ago. It leans on pronouns, on shorthand coined mid-session, and on file names introduced hours back. It continues the thread; it does not…
memory-audit
You are running a memory audit for the user's per-user Claude memory in this project.
context-compression
Use Tokenless for large files and noisy tool outputs before they enter context. Must be used when commands, Read output, diffs, logs, or search results are large.
update-briefs
Updates the briefs/ shared memory directory after a sprint completes. Reads sprint outputs and writes structured summaries for other agents. Use after any sprint completes, or when the user says "update briefs", "sync state", "update project state", or "what changed". Also use automatically after /capture-lessons or…