counsel
Run the weekly observability synthesis on demand and render the coaching brief in the conversation.
Skill by onlooker-community
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/onlooker-community-counsel
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: ml
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/counsel
What's inside
- Skill name
counsel- Description
- Run the weekly observability synthesis on demand and render the coaching brief in the conversation. Reads the onlooker event log, runs a single synthesis pass, writes the brief, and prints it — bypassing the SessionStart staleness gate. Use when the user explicitly invokes /counsel, or wants a fresh improvement brief right now instead of waiting for the next stale-brief regeneration. Supports --show (print the latest brief, no LLM call) and --status.
- License
- MIT
- Instructions
- 139 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Counsel Skill
Counsel's SessionStart hook regenerates a weekly improvement brief only when the
last one has gone stale (`synthesis_interval_days`, default 7) and injects it
invisibly. This skill is the on-demand path: it forces a fresh synthesis right
now and renders the brief into the conversation.
## Setup
Run this once at the start. It sources the plugin helpers, loads config, and
resolves project context.
```bash
set -uo pipefail
PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT:-$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)}"
export CLAUDE_PLUGIN_ROOT="$PLUGIN_ROOT"
source "$PLUGIN_ROOT/scripts/lib/counsel-config.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-events.sh"
source "$PLUGIN_ROOT/scripts/lib/onlooker-project-key.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-project-dir.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-ulid.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-reader.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-synthesize.sh"
source "$PLUGIN_ROOT/scripts/lib/counsel-brief.sh"
REPO_ROOT=$(onlooker_project_repo_root "$(pwd)")
counsel_config_load "$REPO_ROOT"
PROJECT_KEY=$(onlooker_project_key "$(pwd)")
if [[ -z "$PROJECT_KEY" ]]; then
echo "No project key — Counsel needs a git repository (remote or root) to scope briefs. Skipping."
exit 0
fi
BRIEFS_DIR=$(counsel_project_dir "$PROJECT_KEY")
```
## Invocation Modes
### `/counsel` — run the weekly review now (default)
Forces a synthesis pass regardless of brief freshness, writes the brief to
`${ONLOOKER_DIR:-~/.onlooker}/counsel/<project-key>/briefs/<YYYY-WW>.md`, emits
`counsel.brief.generated`, and renders the result. Re-running in the same ISO
week overwrites that week's brief in place. (`$ONLOOKER_DIR` overrides the
storage root; the test suite and non-default installs rely on it.)
```bash
SESSION_ID="${CLAUDE_SESSION_ID:-$(counsel_ulid)}"
export _HOOK_SESSION_ID="$SESSION_ID"
LOOKBACK=$(counsel_config_get '.counsel.lookback_days'); LOOKBACK="${LOOKBACK:-30}"
echo "Running Counsel synthesis over the last ${LOOKBACK} days of events…"
_rc=0
OUTPUT_PATH=$(counsel_generate_brief "$SESSION_ID" "$(pwd)" force) || _rc=$?
if [[ "$_rc" -eq 2 ]]; then
min_events=$(counsel_config_get '.counsel.capture.min_events'); min_events="${min_events:-10}"
echo "Not enough events to synthesize a brief (fewer than ${min_events} in the lookback window). Use the ecosystem long enough to accumulate telemetry, then try again."
exit 0
elif [[ "$_rc" -ne 0 || -z "$OUTPUT_PATH" || ! -f "$OUTPUT_PATH" ]]; then
echo "Counsel synthesis failed. Check that the \`claude\` CLI is on PATH and the onlooker log is readable."
exit 1
fi
```
Then render the brief verbatim to the conversation:
```bash
echo "## Counsel weekly review (\`$(basename "$OUTPUT_PATH" .md)\`)"
echo ""
cat "$OUTPUT_PATH"
echo ""
echo "_Brief saved to ${OUTPUT_PATH}._"
```
### `/counsel --show` — print the latest brief (no LLM call)
Renders the most recent brief already on disk. No synthesis, no events emitted.
```bash
LATEST=$(ls -1 "$BRIEFS_DIR"/*.md 2>/dev/null | sort | tail -1)
if [[ -z "$LATEST" || ! -f "$LATEST" ]]; then
echo "No brief on disk yet for this project. Run \`/counsel\` to generate one."
exit 0
fi
echo "## Counsel brief (\`$(basename "$LATEST" .md)\`)"
echo ""
cat "$LATEST"
```
### `/counsel --status` — brief freshness
Reports the latest brief's age, last-generated time, and path. No LLM call.
```bash
INTERVAL=$(counsel_config_get '.counsel.synthesis_interval_days'); INTERVAL="${INTERVAL:-7}"
LATEST=$(ls -1 "$BRIEFS_DIR"/*.md 2>/dev/null | sort | tail -1)
echo "## Counsel status"
echo "- Project key: ${PROJECT_KEY}"
echo "- Briefs dir: ${BRIEFS_DIR}"
echo "- Stale after: ${INTERVAL} days"
if [[ -z "$LATEST" || ! -f "$LATEST" ]]; then
echo "- Latest brief: none yet (run \`/counsel\`)"
exit 0
fi
if [[ "$(uname)" == "Darwin" ]]; then
mtime=$(stat -f '%m' "$LATEST" 2>/dev/null || echo 0)
human=$(date -r "$mtime" 2>/dev/null || echo "$mtime")
else
mtime=$(stat -c '%Y' "$LATEST" 2>/dev/null || echo 0)
human=$(date -d "@$mtime" 2>/dev/null || echo "$mtime")
fi
now=$(date +%s 2>/dev/null || echo "$mtime")
age_days=$(( (now - mtime) / 86400 ))From onlooker-community/ecosystem, MIT.
More from onlooker-community/ecosystem
- 0.565score
lineage
onlooker-community
Answer "why does this line exist?" — trace a file, or a specific line, back to the change, prompt, agent, and session that produced it.
SkillAFreshness: AClaude Code +9 - 0.565score
cartographer
onlooker-community
Audit CLAUDE.md, AGENTS.md, and .claude/rules/ instruction files for contradictions, stale references, dead rules, scope collisions, and entities no…
SkillAFreshness: AClaude Code +9 - 0.565score
librarian
onlooker-community
Review the librarian's pending memory promotion proposals and lesson candidates queued from past sessions.
SkillAFreshness: AClaude Code +9 - 0.565score
tribunal
onlooker-community
Run a task under multi-agent quality gates.
SkillAFreshness: AClaude Code +9 - 0.565score
warden
onlooker-community
Inspect and control the Warden content gate.
SkillAFreshness: AClaude Code +9 - 0.560score
writing-tests
onlooker-community
How to write tests in the Onlooker ecosystem repo — bats integration tests for hooks and the node:test schema suite.
SkillAFreshness: AClaude Code +9 - 0.555score
list-prompt-rules
onlooker-community
List all prompt rules (global + project) with their pattern, guidance, and per-session fire status.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 11, 2026 (18 days ago)
- Last release
- Oct 24, 2026 (5 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/onlooker-community/ecosystem
- Website
- onlooker.dev
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from onlooker-community
- 0.565score
lineage
onlooker-community
Answer "why does this line exist?" — trace a file, or a specific line, back to the change, prompt, agent, and session that produced it.
SkillAFreshness: AClaude Code +9 - 0.565score
cartographer
onlooker-community
Audit CLAUDE.md, AGENTS.md, and .claude/rules/ instruction files for contradictions, stale references, dead rules, scope collisions, and entities no…
SkillAFreshness: AClaude Code +9 - 0.565score
librarian
onlooker-community
Review the librarian's pending memory promotion proposals and lesson candidates queued from past sessions.
SkillAFreshness: AClaude Code +9 - 0.565score
tribunal
onlooker-community
Run a task under multi-agent quality gates.
SkillAFreshness: AClaude Code +9 - 0.565score
warden
onlooker-community
Inspect and control the Warden content gate.
SkillAFreshness: AClaude Code +9 - 0.560score
writing-tests
onlooker-community
How to write tests in the Onlooker ecosystem repo — bats integration tests for hooks and the node:test schema suite.
SkillAFreshness: AClaude Code +9
Cite this page
counsel skill. Agent Mart. https://agentmart.dev/skills/onlooker-community-counsel (last updated 2026-10-24)