Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/skillberry-ai/cap-evolvenpx agentmods add skills/skillberry-ai/cap-evolve/wikiWrote 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/skillberry-ai/cap-evolve/wiki)<a href="https://agentmods.dev/skills/skillberry-ai/cap-evolve/wiki"><img src="https://agentmods.dev/badge/skills/skillberry-ai/cap-evolve/wiki.svg" alt="Measured on agentmods" 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 | $0.00093 | $0.01665 |
| Opus 5 | $0.00046 | $0.00833 |
| Sonnet 5 | $0.00019 | $0.00333 |
| Haiku 4.5 | $0.00009 | $0.00167 |
Grade A, and why
wiki 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 yesterday.
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 — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
memory-wiki — the weakness graph memory format
This is a memory FORMAT, not a search strategy: any algorithm (hill-climb, gepa, skillopt,
agent-optimize) can select memory_skill: wiki and get this instead of the default
md-files (LEDGER/JOURNAL/PROCESS/INSIGHTS). It replaces per-iteration append-only prose
with a persistent graph of known weaknesses, each carrying its own history of what was
tried and what worked — read once, re-read every iteration, useful across the whole run
instead of scrolling a growing journal.
Absolute path (most important rule)
Write to <run_dir>/wiki/ — the absolute path, never a relative copy inside your
per-iteration working dir. Every iteration gets a fresh working copy; the wiki does not
live inside it and is never copied in or folded back. Writing to the absolute path is what
makes it visible to the next iteration (and to the dashboard's Weakness-graph tab, which
reads wiki/ straight out of the run dir).
The two graphs
- Weaknesses (
wiki/weaknesses/<slug>.md) — what's broken. Persistent across iterations; a weakness'srelatedneighbors are the graph's edges. - Solutions (
wiki/solutions/<weakness-slug>/<sol-id>/) — a kept improvement for one weakness. Every solution[[wikilink]]s back to its weakness.
Weakness node — wiki/weaknesses/<slug>.md
---
slug: tool-call-arg-mismatch
status: in-progress # open | in-progress | completed | solved | reverted
tags: [tool-calling, type-error]
discovered_in_iteration: cand_0003
attacked_in_iterations: [cand_0003, cand_0007]
solved_in_iteration: null
affected_tasks: [task_007, task_011, task_023] # FROZEN after discovery — see below
related:
- slug: schema-drift-after-retry
why: both corrupt the tool-call payload; candidates to merge
solutions:
- "[[tool-call-arg-mismatch-cand_0007]]"
---
# Tool call arg mismatch
## What fails
The agent's tool-call planner passes a dict where the tool expects a string.
## Tasks (found on)
- task_007 — search query sent as JSON object
## References
- `agent/planner.py:88` — builds the args dict; no type coercion before dispatch.
## Rejected Store Memory (RSM)
(Empty at discovery; append dead-end attempts here so a later iteration does not retry them.)
What ships with it
3 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.
- yesterday First seen · 157 lines · 93 tokens per session scan A 08393abf7bbe
wiki is a skill published in the GitHub repository skillberry-ai/cap-evolve (50 stars, last pushed yesterday), licensed Apache-2.0. It adds 93 tokens to every session and 1,665 once invoked, about $0.0005 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 2026-09-05.
Other skills, from other repositories
agent-v3-memory-specialist
Agent skill for v3-memory-specialist - invoke with $agent-v3-memory-specialist.
agent-memory-coordinator
Agent skill for memory-coordinator - invoke with $agent-memory-coordinator.
hindsight-local
Store user preferences, learnings from tasks, and procedure outcomes. Use to remember what works and recall context before new tasks. (user).
session-summaries
What the chat right-panel session summary shows, what it costs, and how to make a session summarize well. Load when the user asks about the session summary panel, why a summary looks wrong or empty, or how to turn it on.
update-setup
One-time setup wizard for the memmy upgrade skill. Triggers: setup update, configure update, 切设置更新, 初始化更新.
skill-creator
Create, edit, improve, tidy, review, audit, or restructure memmy-agent skills and SKILL.md files.