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 sequenzia/agent-alchemy --skill project-learningsgit clone --depth 1 https://github.com/sequenzia/agent-alchemyWrote 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/sequenzia/agent-alchemy/project-learnings)<a href="https://agentmods.dev/skills/sequenzia/agent-alchemy/project-learnings"><img src="https://agentmods.dev/badge/skills/sequenzia/agent-alchemy/project-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/skills/sequenzia/agent-alchemy/project-learnings"><img src="https://agentmods.dev/badge/skills/sequenzia/agent-alchemy/project-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 | $0.00047 | $0.01096 |
| Opus 5.5 | $0.00019 | $0.00438 |
| Sonnet 5.5 | $0.00009 | $0.00219 |
| Haiku 4.5 | $0.00005 | $0.00110 |
Grade A, and why
project-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 1mo 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 — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Learnings
Capture project-specific patterns and anti-patterns into the project's CLAUDE.md. This creates a self-improving feedback loop where discoveries from debugging, development, and review make future Claude sessions smarter.
CRITICAL: Only project-specific knowledge qualifies. Generic programming advice does not belong in CLAUDE.md.
Step 1: Evaluate Discovery
Determine if the finding qualifies as project-specific. The finding must pass at least ONE of these criteria:
| Criteria | Example That Qualifies | Example That Doesn't |
|---|---|---|
| Would a developer unfamiliar with this project likely hit this issue? | "The processOrder() function expects amounts in cents, not dollars" |
"Always validate function inputs" |
| Is this pattern specific to this codebase's architecture, APIs, or conventions? | "The UserProfile type has an optional metadata field that is always present at runtime" |
"Use TypeScript strict mode" |
| Is it something Claude's training data wouldn't cover? | "Never call db.query() without the timeout option — the default is infinite" |
"Use async/await instead of callbacks" |
If NO to all criteria → STOP. Do not add generic programming knowledge to CLAUDE.md. Return to the calling skill and report that no project-specific learning was found.
If YES to any → proceed to Step 2.
Step 2: Read Existing CLAUDE.md
-
Find the project's CLAUDE.md:
- Check the repository root first
- If not found, check if there's a project-level
.claude/directory
-
Parse existing content:
- Understand the existing structure, headings, and conventions
- Look for sections where this learning would fit (e.g., "Known Gotchas", "Bug Patterns", "Conventions", "Known Challenges")
- Check for duplicate or similar entries already present
-
If a similar entry already exists → STOP. Report to the calling skill that this knowledge is already captured. Do not create duplicates.
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.
- 1mo ago First seen · 118 lines · 47 tokens per session scan A e007b351f6c5
project-learnings is a skill published in the GitHub repository sequenzia/agent-alchemy (45 stars, last pushed 4mo ago), licensed MIT. It adds 47 tokens to every session and 1,096 once invoked, about $0.0002 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
performance-audit
Full-stack performance health check across 12 dimensions. Rendering, bundles, assets, API/network, algorithms, memory, database, caching, Web Vitals, backend runtime, concurrency, and framework-specific pathologies. Evidence-based Impact Models with confidence tiers and a prioritized optimization roadmap. Switches…
evolving-status
View Self-Evolving Loop session status, history, and memory metrics.
ccc-knowledge
Search, browse and add lessons in the local CC Commander knowledge base: recent lessons, topic search, category view, stats and file-edit captures. Use when the user asks what we learned, wants past lessons, the knowledge base, or types $ccc-knowledge.
ccc-handoff
Write a dense handoff file and tell the user to resume in a new chat before a long context degrades output quality. Use when a chunk of work is done, the topic is about to switch, the model starts forgetting context, or a long orchestration run is mid-flight.
ccc-memory
Keep durable project memory in two tiers: CLAUDE.md for always-loaded facts and a memory/ directory of dated notes, decisions, gotchas and preferences loaded on demand. Use when recording a decision, root cause, correction or preference future sessions must keep.
ccc-resume-session
Load a saved session file from /.claude/sessions/ (latest, by date, or by path), read it fully, and brief project, state, what not to retry, blockers and next step before touching anything. Use when resuming work, continuing where you left off, or after a fresh session.