Standup Context

Data-gathering scripts for sync/standup reviews, reducing tool calls from ~15+ down to 2-3. Collects git state, canonical docs, telemetry, and cross-session memory into structured markdown for synthesis.

Capabilities

This skill provides two data-gathering capabilities via shell scripts. All script paths are relative to this skill's directory.

1. Git and Canonical Docs

When to use: Always. This is the primary data source for any standup or status review.

What it collects:

• Git: recent commits (12h), status, diff stat, branches — always collected
• Canonical docs: file listing with titles from caller-specified directories
• Status reports: 3 most recent craft-status files (content)
• Learnings: cross-cutting entries, per-charter/plan learnings, recent ADRs
• Waste snake: recent observations, total count, latest ledger entry
• Memory: cross-session context (first 200 lines)

Each non-git section is controlled by an env var. Sections are skipped when their var is unset, so the script collects only what the caller tells it about. This keeps the script decoupled from any specific project layout.

Env vars:

Var	Purpose	Example
CANONICAL_ROOT	Parent dir for canonical doc subdirs	per /docs/layout
CANONICAL_DIRS	Space-separated subdirs to scan	roadmaps charters backlogs plans
REPORTS_DIR	Directory with craft-status-* files	per /docs/layout
LEARNINGS_FILE	File with cross-cutting L* entries	per /docs/layout
LEARNINGS_DIRS	Space-separated dirs to scan for ## Learnings sections	subdirs under CANONICAL_ROOT
ADR_DIR	Directory with ADR .md files	per /docs/layout
WASTE_SNAKE	Path to waste snake file	per /docs/layout
MEMORY_FILE	Path to cross-session memory file	~/.claude/projects/.../memory/MEMORY.md

How to run: Use /docs/layout to resolve all artifact paths, then pass them as env vars to the script. The command outputs KEY=value lines that map directly to the env vars below.

# 1. Resolve paths via /docs/layout (outputs KEY=value for all keys)
# 2. Pass non-empty values as env vars:
CANONICAL_ROOT=<from /docs/layout> \
CANONICAL_DIRS="roadmaps charters backlogs plans" \
REPORTS_DIR=<from /docs/layout> \
LEARNINGS_FILE=<from /docs/layout> \
LEARNINGS_DIRS="<CANONICAL_ROOT>/charters <CANONICAL_ROOT>/plans" \
ADR_DIR=<from /docs/layout> \
WASTE_SNAKE=<from /docs/layout> \
MEMORY_FILE=~/.claude/projects/-Users-me-projects-my-repo/memory/MEMORY.md \
bash <SKILL_DIR>/scripts/gather-git-and-docs.sh

Omit any env var whose /docs/layout value is empty — the script skips that section.

2. Telemetry

When to use: Always attempt this. It gracefully outputs "TELEMETRY UNAVAILABLE" when credentials are missing, so it is safe to run unconditionally.

What it collects:

• session_overview — sessions, tokens, cost
• agent_usage_daily — agent invocations
• cost_by_agent — cost breakdown per agent/model
• skill_frequency — skill activations

Requires TB_READ_TOKEN (or TB_TOKEN) and TB_HOST env vars. Auto-sources .env.prod from repo root if env vars are not set.

How to run: Execute ./scripts/gather-telemetry.sh relative to this skill's directory, passing the current project name.

bash <SKILL_DIR>/scripts/gather-telemetry.sh "$(basename "$(pwd)")"

Path Resolution

<SKILL_DIR> is the directory containing this SKILL.md file. When loading this skill, resolve the absolute path to this file and use its parent directory as the base for all script paths. For example, if this file was loaded from /Users/me/projects/my-repo/skills/standup-context/SKILL.md, then the scripts are at /Users/me/projects/my-repo/skills/standup-context/scripts/.

Testing

bash <SKILL_DIR>/scripts/gather-git-and-docs.test.sh
bash <SKILL_DIR>/scripts/gather-telemetry.test.sh