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 traviswye/ContextAtlas --skill refresh-atlasgit clone --depth 1 https://github.com/traviswye/ContextAtlasWrote 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/traviswye/contextatlas/refresh-atlas)<a href="https://agentmods.dev/skills/traviswye/contextatlas/refresh-atlas"><img src="https://agentmods.dev/badge/skills/traviswye/contextatlas/refresh-atlas/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/traviswye/contextatlas/refresh-atlas"><img src="https://agentmods.dev/badge/skills/traviswye/contextatlas/refresh-atlas.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 | $0.00101 | $0.01351 |
| Opus 5.5 | $0.00040 | $0.00540 |
| Sonnet 5.5 | $0.00020 | $0.00270 |
| Haiku 4.5 | $0.00010 | $0.00135 |
Grade A, and why
refresh-atlas 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
refresh-atlas — re-extract the pending sources
contextatlas index --structural (run by the git hooks) keeps the
local cache's symbols and git signal current and records every source
whose claims are out of date in .contextatlas/local/pending.json.
This skill extracts claims for those sources, and only those, and
writes them into .contextatlas/atlas.json.
It may run headless (claude -p, started by contextatlas refresh-claims). Never ask the user a question and never wait for
input. When something blocks you, stop and print one line starting
with REFRESH-FAILED: that says what blocked you; do not leave a
half-written atlas.json behind.
Arguments: $ARGUMENTS (for example --max-sources 25), passed to
list-extraction-sources below. Empty means no cap.
This skill reuses /index-atlas. Read
.claude/skills/index-atlas/SKILL.md once, at the start, for:
- "Canonical atlas schema (v1.4)" and "Schema invariants": the exact shape every claim, key and top-level field must have;
- "Phase B" steps 1 to 3: how to extract a claim set from one ADR, one
docstring entry and one commit, and how to fill each claim's
source,source_path,source_sha,symbol_idsandsymbol_candidates; - "Phase C": the gate commands and how to react to each failure.
Everything below overrides /index-atlas where they differ.
Step 1 — List the pending sources
contextatlas list-extraction-sources --pending $ARGUMENTS --output .contextatlas/local/refresh-sources.json
A non-zero exit is a setup failure: print REFRESH-FAILED: with its
stderr and stop.
Read .contextatlas/local/refresh-sources.json. Its pending section
says how many sources are queued (queued), how many this manifest
covers (selected) and how many stay for a later run (remaining).
If pending.selected is 0, print REFRESHED nothing pending and stop
without touching atlas.json.
Then Read .contextatlas/prompts/extraction.md: that file's content is
EXTRACTION_PROMPT, the frozen prompt. If it is missing, print
REFRESH-FAILED: .contextatlas/prompts/extraction.md is missing; run contextatlas init and stop. Never write a prompt of your own.
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 · 123 lines · 101 tokens per session scan A f54761f437da
refresh-atlas is a skill published in the GitHub repository traviswye/ContextAtlas (7 stars, last pushed yesterday), licensed MIT. It adds 101 tokens to every session and 1,351 once invoked, about $0.0004 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 2026-10-15.
Other skills, from other repositories
decision-memory
Use when planning, designing, reviewing, or changing code in a repository that keeps ADRs (usually docs/adr) — to load the decisions that already govern the work, check a plan or diff against them, or record a new decision. Also use when a choice feels already-settled and you cannot find where it was settled.
memory-continuity
Preserves cross-session continuity with deterministic MEMORY.md, index.md, and snapshot handoffs under .agentic/memories/.
plan-memory-index
Maintains concise indexed memory entries for applied plans under .agentic/memories/ with clear boundaries between memory and plan sources. Memory retrieval is prioritized before loading historical plan files.
plan
Structured technical planning before implementation — architecture and task breakdown.
architecture-governance
Keep code changes conformant to a repo's real, declared architecture using ArchSteer. Use when starting work in an unfamiliar repo, before finishing/committing an architecturally significant change (new dependency, new data access, new external call, new layer), or when asked about a codebase's architecture, layers…
mpm-adr
Architecture Decision Records (ADRs) — opt-in convention for documenting significant, hard-to-reverse architectural decisions using the Nygard template.