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 linkpranay-ai/context-engineering-protocol --skill ult-cep-retrofitgit clone --depth 1 https://github.com/linkpranay-ai/context-engineering-protocolWrote 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/linkpranay-ai/context-engineering-protocol/ult-cep-retrofit)<a href="https://agentmods.dev/skills/linkpranay-ai/context-engineering-protocol/ult-cep-retrofit"><img src="https://agentmods.dev/badge/skills/linkpranay-ai/context-engineering-protocol/ult-cep-retrofit/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/linkpranay-ai/context-engineering-protocol/ult-cep-retrofit"><img src="https://agentmods.dev/badge/skills/linkpranay-ai/context-engineering-protocol/ult-cep-retrofit.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 | $0.00084 | $0.03235 |
| Opus 5 | $0.00042 | $0.01618 |
| Sonnet 5 | $0.00017 | $0.00647 |
| Haiku 4.5 | $0.00008 | $0.00324 |
Grade A, and why
cep-retrofit 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 10d 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 — 215 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CEP Retrofit
Overview
Every CEP consumption contract (CONSUMING-CONTEXT-PACKAGE.md, CONSUMING-COMPILED-GUIDELINES.md,
CONSUMING-CODE-GRAPH.md) already says it's meant to be "referenced by a one-line pointer from
each consuming skill's SKILL.md/.prompt.md rather than copied into each one." That's cheap in
principle. In practice, adding that pointer to an existing skill library — correctly, one skill
at a time, in a format that matches the library's own conventions — takes someone who has read
PROTOCOL.md and the relevant contract(s) end to end. For a library with a dozen or more skills,
that friction is real, and it sits squarely between CEP and "every retrofitted library becomes a
live adoption story for free."
This skill inventories an existing skill library — any shape, seen or unseen — asks a human which
skills should consume which contract(s), then drafts the minimal pointer insertion, previews every
change, and writes only what's confirmed. It invents no new consumption mechanism. It reuses
the three CONSUMING-*.md contracts verbatim; if a real retrofit need doesn't map to any of them,
that's a gap in CEP itself to raise separately (see "What this skill deliberately does not do").
The one hard rule: zero hardcoded knowledge of any specific library
This skill must run identically against a library it has never seen, in a format nobody has described to it yet, and against the author's own private skill collections. Concretely:
- No real library's name, skill name, or path fragment ever appears in this skill's own
instructions, its helper script, or its tests — not even as "for example, in a library like...".
Every example anywhere in this skill is a fabricated placeholder (
example-skill/,widget-reviewer/). - Detection heuristics are shape-based, with one narrow, documented exception: a small set of
root-level filenames (
AGENTS.md,README.md, etc. — a broadly used, cross-project convention, not any specific library's own name) are excluded by literal name, but only at the target root itself, never at any nested depth. "Does this directory contain a file matchingSKILL.md/skill.md" is fine; "if the file is literally namedX" for any real, privateXis not. scripts/cep_retrofit.pyowns every heuristic that can be made deterministic (inventory, description extraction, recommendation signals, idempotency checking, insertion-point detection) so those rules live in one tested place, not re-derived ad hoc per run. All ledger- style mechanical work goes through the script — never hand-roll a directory walk or a frontmatter regex inline. See the script's own module docstring for the exact subcommand surface.
What ships with it
2 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.
- 10d ago First seen · 215 lines · 84 tokens per session scan A 46813afd1de5
cep-retrofit is a skill published in the GitHub repository linkpranay-ai/context-engineering-protocol (9 stars, last pushed yesterday), licensed Apache-2.0. It adds 84 tokens to every session and 3,235 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
curate-a-team-library
Use when building a managed team skills library for a real stack. Map work to shelves, browse before curating, write meaningful whyHere notes, and create a starter pack once the first pass is solid.
migrate-skills-between-libraries
Use when moving skills between library workspaces or upgrading from a personal library to a team library. Export from one workspace, import into another.
audit-library-health
Use when checking the overall health of a skills library. Run doctor, validate, check for stale skills, and verify generated docs are in sync.
build-workspace-docs
Use when regenerating README.md and WORKAREAS.md in a managed library workspace. Always dry-run first to preview changes.
share-a-library
Use when a managed library is ready to publish to GitHub and hand to teammates as an install command. Run the GitHub publishing steps, then return the exact shareable install command.
update-installed-skills
Use when syncing or updating previously installed skills to their latest version. Always dry-run updates before applying, and check for breaking changes.