跳到正文
cccskills
免费在 GitHub 公开

signals

Read HQ-native company signals such as decisions, actions, risks, commitments, and wins.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core星标 90(GitHub 90 + cccskills 0)

实测评分 未实测

查看安装方法

包含的文件(2)

  • SKILL.md4.9 KB
  • agents/openai.yaml119 B

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

/signals — Read HQ-native signals (decisions, action items, wins, risks…)

Signals are an HQ-native, per-company store of meeting/comms intelligence — discrete, cited items extracted from meeting notes and other sources. This is the generic, multi-company reader. A company may also ship its own namespaced interface (e.g. {co}:signals, {co}:action-items) over the same store; this skill serves all companies.

Canonical store: companies/{co}/signals/{type}/{sha256}.md (company-wide) and companies/{co}/signals/@{audienceKey}/{type}/{sha256}.md (visible only to the people privy to the source — spec core/knowledge/public/hq-core/ontology-local-spec.md), daily index companies/{co}/signals/_index/{YYYY-MM-DD}.json. Mechanism/schema detail (read once if needed): core/knowledge/public/hq-core/native-knowledge-stores.md.

Signal types (8)

TypeUse it for
action_itemtasks/to-dos assigned to an owner
commitmentexplicit promises a participant made
decisiondecisions reached
riskblockers, concerns, threats raised
questionopen questions, unresolved
key_pointnotable discussion points
participant_contributionwhat a specific person contributed
summaryper-meeting summaries

"Wins / accomplishments" → read decision + key_point (and summary). "Follow-ups" → action_item + commitment.

Each signal .md has frontmatter: canonical_content (one-line normalized statement), citations[] (location timestamp, source_ref, verbatim text), entity_refs[] (person/…, project/…), source_ref, type, signal_id.

Hard rule

HQ-native first. Read the signals store to answer these asks. Never default to scraping email or a third-party notetaker. If the store is empty/unactivated, steer the user to activation (below) — don't fabricate or guess.

Resolution (run every time)

  1. Resolve company. bash core/scripts/hq-session.sh get company_slug; else infer from cwd/manifest; else ask.
  2. Load preference (company over global). Read companies/{co}/settings/knowledge/preferences.yaml then personal/settings/knowledge-preferences.yaml. Built-in default: signals_enabled: true. If a company has signals_enabled: false, tell the user it's disabled and offer to enable it (flip the flag) rather than reading.
  3. Classify the ask via the type table (and any aliases.signals in the preference file): action items → action_item, "what did we decide" → decision, wins → decision+key_point, risks → risk, open questions → question, "{person} committed" → commitment filtered by entity_refs person.

Reading

  • Recency: use the latest _index/{date}.json files (newest dates first) to find recently-written signals; or list newest files under both companies/{co}/signals/{type}/ and companies/{co}/signals/@*/{type}/. Skip _candidates/ (not yet gardened). Mark each scoped item as scoped when presenting it, and never repeat a scoped item into a shared surface (a channel post, a company doc).
  • Filter: by type (dir), by owner/person (entity_refs contains person/{slug}), by project (project/{slug}), by date.
  • Present: group by type; show canonical_content per item with its owner/source; cite the meeting (source_ref) and timestamp from citations[] when useful. For mutable action-item tracking with status lifecycle, note that a company may ship a dedicated namespaced tracker (e.g. {co}:action-items); this skill is read-only.

Empty / not-activated branch

If companies/{co}/signals/ is empty or absent, signals haven't been extracted for this company yet. Surface the turnkey ladder (full version in the knowledge doc + USER-GUIDE):

  1. /designate-team {co} to make the company cloud-backed.
  2. Capture meetings with the HQ meeting bot (/meeting-notes covers this) → notes ingest.
  3. Signals are extracted from ingested notes into companies/{co}/signals/ automatically once provisioned.
  4. Note: billing is live. On the legacy plans, inviting the HQ meeting bot needs the company on HQ Workforce (a Starter company gets "HQ Workforce plan required to invite a meeting bot" with an upgrade link; the owner upgrades with hq billing upgrade --company {co} or on the console Billing page). On the usage card (new signups from 2026-10-16), relay what HQ returns. Signals extraction runs on HQ cloud over the notes that arrive. Plan details: core/knowledge/public/hq-core/plans-and-pricing.md.

Output

Quiet, plain, grouped summary of the signals the user asked for — decisions, action items, wins, risks, etc. — with light citations.

See also

  • /meeting-notes — the source meeting notes
  • /ontology — entities behind the signals

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

accept

免费

Accept HQ membership invites from a magic link or raw token.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

adr

免费

Capture qualifying architecture decisions as ADRs in the right HQ or repo location.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

architect

免费

Find shallow modules and propose deepening refactors, ranked with file refs; never edits code. Use when code is hard to change or planning a refactor.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

Generate dark ASCII banner graphics for HQ and social launch assets.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

Compare approaches to best-in-class prior art, merge existing work, verify cruft removal, and recommend the most durable design before PRD work.

中文简介正在准备中。当前显示原文说明。

indigoai-us/hq-core902026年10月26日 更新

查看 indigoai-us 的全部技能

报告此技能的问题