composition-evaluate SKILL.md skillA
composition-evaluate from jmagly/aiwg is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add jmagly/aiwg/composition-evaluate@git:20260822.2a3b34aPins git:20260822.2a3b34a in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- jmagly/aiwg
- Path
agentic/code/addons/composition-engine/skills/composition-evaluate/SKILL.md- Versions recorded
- 1
- Latest version
- git:20260822.2a3b34a, committed Aug 22, 2026
- Size
- 1,500 bytes
- Repository stars
- 229
- License
- MIT
Compare FlowGraph composition policies with fixed tasks, requested-versus-realized resources, failure injection, and an evidence claim gate.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Composition evaluation Use this skill when a composition policy needs reproducible comparison against a single-pass baseline. 1. Start from `benchmarks/composition-policy-benchmark.v1.json` and preserve fixed tasks, settings, budgets, seeds, metrics, and thresholds. 2. Run `aiwg composition benchmark <manifest.json>`; use `--raw-out` and `--summary-out` to retain both evidence layers. 3. Compare success-conditioned cost and latency, not unconditioned cheap failures. Review speed-of-accuracy and strict-LCM-versus-adaptive deltas. 4. Require an independent or human evaluation path and inspect self-judge bias. 5. Review every failure-injection outcome and recovery receipt. 6. Keep synthetic conformance distinct from provider evidence. Do not open the claim gate without repeated trusted runs, independent evaluation, confidence intervals, and task-family replication. Composition graphs remain `flow.aiwg.io/v1alpha1` `FlowGraph`; do not introduce a fourth-level DNS API group. Do not request or persist private chain-of-thought. @implements #2118
Read the whole file at its exact version.
How to install
mdr add jmagly/aiwg/composition-evaluate@sha256:73a900ce6e0a1cf8The line at the top of this page pins the label git:20260822.2a3b34a, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_464wooagsnwduvu7)
3 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Aug 22, 2026.
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (1500 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
jmagly/aiwg · 229 stars · license MIT · pushed 2026-11-07 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_464wooagsnwduvu7 GET https://markdownregistry.com/api/v1/resolve?ref=jmagly/aiwg/composition-evaluate GET https://markdownregistry.com/api/v1/blob/73a900ce6e0a1cf848309ff3d3a6669a43b02400fa64aa535516ed2c88a4606e
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.