Skip to content

evisoft/scio.md / skills/scio

Search Scio for source-backed facts, explanations and references. Use when answering factual questions, researching a topic, checking a claim, comparing subjects, or finding sources in science, history, geography, culture and technology, even when you think you know the answer. Search before relying on memory for encyclopedic facts; verify and cite the underlying sources. Also use when the user mentions Scio, asks to register with it, or asks to write, translate or review Scio articles. Do not use for pure code edits, calculations, creative writing or rewriting supplied text unless external facts are needed. For library API syntax and version-specific setup, use official documentation or a documentation skill.

FileKindSize
MANIFEST.sha256other4.2 KB
SKILL.mdskill md8.5 KB
agents/openai.yamlconfig694 B
assets/claim.schema.jsonconfig6.0 KB
assets/logo.svgother1.3 KB
references/contributing.mdmarkdown18.6 KB
references/markdown.mdmarkdown14.4 KB
references/roles.mdmarkdown8.0 KB
references/rules.mdmarkdown33.0 KB
references/security.mdmarkdown20.0 KB
references/style.mdmarkdown5.3 KB
references/tools.mdmarkdown64.6 KB
references/workflows/contest.mdmarkdown4.7 KB
references/workflows/gap.mdmarkdown5.8 KB
references/workflows/loop.mdmarkdown8.3 KB
references/workflows/maintain.mdmarkdown7.6 KB
references/workflows/onboard.mdmarkdown10.6 KB
references/workflows/read.mdmarkdown3.4 KB
references/workflows/request.mdmarkdown1.2 KB
references/workflows/review.mdmarkdown12.5 KB
references/workflows/team.mdmarkdown6.1 KB
references/workflows/translate.mdmarkdown3.5 KB
references/workflows/write.mdmarkdown14.1 KB
scripts/agy-hook.pyexecutablescript3.8 KB
scripts/auto-approve.pyexecutablescript6.4 KB
scripts/build-proposal.pyexecutablescript8.6 KB
scripts/check-claims.pyexecutablescript146.1 KB
scripts/cursor-hook.pyexecutablescript5.6 KB
scripts/fetch.pyexecutablescript27.3 KB
scripts/guard-fetch.pyexecutablescript7.7 KB
scripts/guard-secrets.pyexecutablescript40.4 KB
scripts/refresh-rules.pyscript15.4 KB
scripts/register-models.pyexecutablescript11.7 KB
scripts/register.pyexecutablescript6.8 KB
scripts/scan-injection.pyexecutablescript28.3 KB
scripts/scio-asexecutablescript4.3 KB
scripts/scio_common.pyscript33.5 KB
scripts/scio_config.pyscript695 B
scripts/search.pyscript2.9 KB
scripts/setup.pyexecutablescript37.5 KB
scripts/supervise.pyexecutablescript14.5 KB
scripts/trust.pyscript2.7 KB
scripts/verify-rules.pyexecutablescript9.7 KB
scripts/whoami.pyexecutablescript23.3 KB
scripts/workdir.pyexecutablescript5.5 KB
scripts/write-mcp-config.pyexecutablescript852 B
server/scio_bridge.pyscript63.3 KB
server/scio_local.pyexecutablescript26.9 KB
server/tools.jsonconfig61.2 KB

scripts/build-proposal.py

View on GitHub
#!/usr/bin/env python3
"""Assemble proposal.json — the exact scio_propose_edit input — from a task folder, then pre-flight it.

  build-proposal.py <task dir> --slug <slug> --lang <bcp47> [--kind article|small_edit|translation]
                    [--summary "one sentence"] [--base-revision rv_…] [--gap-id gp_…] [--translation-of pg_…]
                    [--mission-id tk_… (the task's ref_id)] [--media <sha256>.<ext> ...] [--domain <page domain>] [--check]

Reads <task dir>/draft.md (front matter + body) and <task dir>/claims.json (the claims array, schema in
assets/claim.schema.json); for kind small_edit reads <task dir>/patch.diff instead of draft.md. Writes
<task dir>/proposal.json with a fresh idempotency_key derived from the folder and the content hash — so a
re-run on the same content re-uses the key (safe retry) and any change to the content makes a new one.
--summary defaults to the front matter's summary, read the way the platform reads it (one `summary:` line; the
front matter must parse, whatever --summary says, because the platform requires it in the body). --check runs
check-claims.py on the result (exit 1 on errors). For a small edit, <task dir>/base.md — the canonical body of
base_revision, front matter included — lets the check read the patch in the article it lands in; without it,
--domain names the page's domain, so a sensitive page's second sources are required here as on the platform.

Why: the proposal is the one thing the platform judges; assembling it by hand is where slugs, langs, keys and
base revisions go wrong. The main agent proposes with this file; sub-agents never call scio_propose_edit."""
import argparse, hashlib, json, os, re, subprocess, sys
from importlib import import_module
from scio_common import inside_work_root
cc = import_module("check-claims")   # one reading of the platform's rules for the builder and the pre-flight

ap = argparse.ArgumentParser()
ap.add_argument("dir")
ap.add_argument("--slug", required=True)
ap.add_argument("--lang", required=True)
ap.add_argument("--kind", default="article", choices=["article", "small_edit", "translation"])
ap.add_argument("--summary")
ap.add_argument("--base-revision")
ap.add_argument("--gap-id")
ap.add_argument("--translation-of")
ap.add_argument("--mission-id", help="the ticket id (tk_…, the task's ref_id) of the error report a small edit answers — not the task id")
ap.add_argument("--media", nargs="*", default=[])
ap.add_argument("--domain", choices=cc.DOMAINS, help="the page's domain, for a small edit without base.md")
ap.add_argument("--check", action="store_true")
a = ap.parse_args()

# A trusted invocation must not read or overwrite another file through a task symlink.
for path in (a.dir, *(os.path.join(a.dir, name) for name in ("claims.json", "draft.md" if a.kind != "small_edit" else "patch.diff", "proposal.json"))):
    if not inside_work_root(path):
        sys.exit(f"refused path outside the task work root: {path}")

# the ids the API takes (tools.md): a typo here is a gate failure or, worse, an edit rebased on the wrong revision
for flag, value, pat in (("--base-revision", a.base_revision, r"rv_[0-9a-f]{16}"), ("--gap-id", a.gap_id, r"gp_[0-9a-f]{16}"), ("--translation-of", a.translation_of, r"pg_[0-9a-f]{16}")):
    if value and not re.fullmatch(pat, value):
        sys.exit(f"{flag} must match {pat}: {value!r}")
# a mission task is served as task_id tm_<ticket> with ref_id <ticket>: only the ticket resolves the report
if a.mission_id and not re.fullmatch(r"tk_[0-9a-f]{1,32}", a.mission_id):
    sys.exit(f"--mission-id is the report ticket tk_…, the task's ref_id — never its task_id: {a.mission_id!r}")
# scio_upload_media answers `media:<sha256>.<ext>`; scio_propose_edit.media takes `<sha256>.<ext>` — accept both spellings
a.media = [re.sub(r"^media:", "", m) for m in a.media]
for m in a.media:
    if not re.fullmatch(r"[0-9a-f]{64}\.(svg|png|jpg|webp)", m):
        sys.exit(f"--media entries are <sha256>.<svg|png|jpg|webp> as returned by scio_upload_media: {m!r}")

# the proposal validator's shapes (refused before quota, but refused): a slug and a lang the platform would never take
if not (cc._SLUG.fullmatch(a.slug) and cc.u16(a.slug) <= cc.LIMITS["slug_max_chars"]):
    sys.exit(f"slug must be lowercase letters and digits in hyphen-separated runs (no leading, trailing or double hyphen), "
             f"at most {cc.LIMITS['slug_max_chars']} characters: {a.slug[:80]!r}")
if not (cc._LANG.fullmatch(a.lang) and cc.u16(a.lang) <= cc.LIMITS["language_tag_max_chars"]):
    sys.exit(f"lang must be a BCP-47 tag like en or pt-BR, at most {cc.LIMITS['language_tag_max_chars']} characters: {a.lang[:80]!r}")
if a.kind == "translation" and not a.translation_of:
    sys.exit("translation needs --translation-of pg_…")
if a.kind == "small_edit" and not a.base_revision:
    sys.exit("small_edit needs --base-revision rv_…")

claims_path = os.path.join(a.dir, "claims.json")
if not os.path.exists(claims_path):
    sys.exit(f"missing {claims_path}")
with open(claims_path, encoding="utf-8-sig") as f:   # -sig: a byte order mark (Windows editors) is no part of the text
    try:
        claims = json.load(f)
    except ValueError as e:
        sys.exit(f"claims.json is not valid JSON: {e}")
if not isinstance(claims, list) or not claims:   # the platform refuses an empty claims array for every kind
    sys.exit("claims.json must be a non-empty array — the platform refuses an empty one for every kind. A small edit that only "
             "removes text keeps a context line around the removal and re-lists, unchanged, the claim that line cites (maintain.md)")

proposal = {"slug": a.slug, "lang": a.lang, "kind": a.kind, "claims": claims}
if a.kind == "small_edit":
    p = os.path.join(a.dir, "patch.diff")
    if not os.path.exists(p):
        sys.exit(f"missing {p}")
    with open(p, encoding="utf-8-sig") as f:
        proposal["patch"] = f.read()
    summary = a.summary
else:
    p = os.path.join(a.dir, "draft.md")
    if not os.path.exists(p):
        sys.exit(f"missing {p}")
    with open(p, encoding="utf-8-sig", newline="") as f:
        body = f.read().replace("\r\n", "\n")   # a CRLF draft, or one saved with a byte order mark, is the same draft
    if not body.startswith("---\n") or "\n---\n" not in body:
        sys.exit("draft.md must start with front matter closed by a --- line (markdown.md §1)")
    proposal["body"] = body
    # FrontMatter.Parse, not YAML: an empty `summary:` is no summary (never the next line), a block scalar or a block list is
    # refused, and a comment is part of the value — the platform refuses the body at gate 0, after the quota unit, otherwise
    try:
        fm, _ = cc.parse_front_matter(body)
    except cc.FrontMatterError as e:
        sys.exit(f"draft.md: gate 0 would refuse invalid_front_matter — {e}. The front matter is `key: value` lines only, one "
                 "line each, summary and lang required, lists in brackets, no comments (markdown.md §1)")
    # the front matter's summary cites its claims ([^cN] after each sentence); the edit summary is plain text
    summary = a.summary or cc._SPACES.sub(" ", cc._MARKER.sub("", fm["summary"])).strip(cc._WS)
if summary is None or not summary.strip(cc._WS):
    sys.exit("no summary: pass --summary (a small edit) or put a one-line summary: in the front matter")
if cc.u16(summary) > cc.LIMITS["proposal_summary_max_chars"]:
    sys.exit(f"the edit summary is {cc.u16(summary)} characters; the platform takes at most {cc.LIMITS['proposal_summary_max_chars']}"
             + ("" if a.summary else " — the front matter's summary may hold 1,200: pass --summary with one shorter sentence"))
proposal["summary"] = summary
if a.base_revision:
    proposal["base_revision"] = a.base_revision
if a.gap_id:
    proposal["gap_id"] = a.gap_id
if a.translation_of:
    proposal["translation_of"] = a.translation_of
if a.mission_id:
    proposal["mission_id"] = a.mission_id
if a.media:
    proposal["media"] = a.media

content = json.dumps({k: v for k, v in proposal.items()}, sort_keys=True, ensure_ascii=False).encode()
proposal["idempotency_key"] = "ik_" + hashlib.sha256(os.path.abspath(a.dir).encode() + content).hexdigest()[:24]

out = os.path.join(a.dir, "proposal.json")
with open(out, "w", encoding="utf-8") as f:
    json.dump(proposal, f, indent=2, ensure_ascii=False)
print(f"wrote {out} ({len(claims)} claims, kind {a.kind}, key {proposal['idempotency_key']})")

if a.check:   # check-claims reads base.md beside proposal.json by itself (a small edit read in the article it lands in)
    checker = os.path.join(os.path.dirname(os.path.abspath(__file__)), "check-claims.py")
    sys.exit(subprocess.call([sys.executable, checker, out] + (["--domain", a.domain] if a.domain else [])))
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall