forcedotcom/sf-skills / skills/sales-call-scoring-configure
sales-call-scoring-configure
VerifiedConfigure, set up, or repair AI Call Scoring (a.k.a. Call Coaching / Coaching Competencies) on an ECI-enabled Salesforce org. Flips ECICallScoringEnabled via SOAP Metadata API, enables Einstein Generative AI Platform if off, and manages Coaching Competency records (EnablementCompetencyDef Tooling API): install the 6 OOTB set, the 7 Momentum best-practice prompts, or conversationally author/edit/activate/deactivate custom competencies (8-active + 16000/255-char caps). TRIGGER when: user wants to turn on AI Call Scoring or Call Coaching, install OOTB or best-practice competencies, create/edit custom competencies, re-enable after sandbox refresh, or verify/repair partial Call Coaching setup on a 68+ ECI org. DO NOT TRIGGER when: no ECI license (direct to Agentforce for Sales), user wants to enable ECI/Momentum themselves (out of scope — Setup), Pipeline Management (use sales-agentforce-pipeline-management-configure), or custom agent (use agentforce-generate).
scripts/enable-call-scoring.sh
View on GitHub#!/bin/bash
# Enable Call Coaching (a.k.a. AI Call Scoring / Coaching Competencies) on a
# Salesforce org via SOAP Metadata API.
#
# Handles the PRD-scoped org states for the ECI arm:
# 1. Not available → refuse with PRD-verbatim message.
# 2. Available, not enabled → enable Einstein GenAI if off, flip
# enableECICallScoring.
# 3. Enabled, no competencies → exit 0 with "ready for competency setup".
# 4. Enabled, competencies exist → exit 0 with "already enabled, N competencies".
#
# River Rush (Momentum) arm detection is deferred until that arm goes live —
# see SKILL.md. Re-add it as a 5th state / advisory when it does.
#
# Usage: ./enable-call-scoring.sh <org-alias>
set -euo pipefail
export NO_COLOR=1
export FORCE_COLOR=0
ORG_ALIAS="${1:-}"
if [[ -z "$ORG_ALIAS" ]]; then
echo "Error: Missing org alias"
echo "Usage: $0 <org-alias>"
exit 1
fi
echo "Enabling Call Coaching for org: $ORG_ALIAS"
echo ""
# ---------------------------------------------------------------------------
# Auth
# ---------------------------------------------------------------------------
SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
# shellcheck source=shared/auth.sh
source "$SCRIPT_DIR/shared/auth.sh"
# SOAP Metadata API embeds the session token in the request body (see
# soap_read below). This is the only caller in the skill that still needs
# the raw ACCESS_TOKEN in shell scope — everything else goes through
# `sf api request rest`. See shared/auth.sh header for the rationale.
resolve_org_auth_with_token "$ORG_ALIAS" || exit 1
# Deliberately no "Org ID: ..." / "Instance URL: ..." banner here — those are
# admin-comms leaks (rule #5). The org alias is already the admin's mental
# handle; the record Id and pod URL belong in --verbose debug output only.
# ---------------------------------------------------------------------------
# Phase 0: ECI purchase gate — ConversationPilot OrgPermission
# ---------------------------------------------------------------------------
# ConversationPilot reflects whether the org is licensed for Einstein
# Conversation Insights at all, independent of whether an admin has flipped
# enableCallCoaching on in Setup. Checking it first lets us tell "you don't
# have ECI, go buy it" apart from "you have ECI, just turn it on" instead of
# inferring a license gap from a readMetadata fault on enableCallCoaching.
#
# Some org shapes (older sandboxes, non-EinsteinConversationInsights-provisioned
# orgs) reject the OrganizationSettingsDetail query with INVALID_TYPE. That's
# not a licensing signal — it's a metadata-availability signal. Treat any
# non-conclusive result as "unknown, defer to SOAP" and only refuse when the
# query explicitly comes back "false".
CONVERSATIONPILOT_ENABLED=""
if CONVERSATIONPILOT_RAW=$(sf data query --target-org "$ORG_ALIAS" --use-tooling-api \
--query "SELECT SettingValue FROM OrganizationSettingsDetail WHERE SettingName='ConversationPilot' AND DataType='OrgPermissionSettings'" \
--json 2>/dev/null); then
CONVERSATIONPILOT_JSON=$(echo "$CONVERSATIONPILOT_RAW" | awk '/^{/,EOF')
if echo "$CONVERSATIONPILOT_JSON" | jq -e '.status == 0' >/dev/null 2>&1; then
CONVERSATIONPILOT_ENABLED=$(echo "$CONVERSATIONPILOT_JSON" | jq -r '.result.records[0].SettingValue | select(. != null)')
fi
fi
if [[ "$CONVERSATIONPILOT_ENABLED" == "false" ]]; then
echo "Call Coaching couldn't be turned on because this org doesn't have Einstein Conversation"
echo " Insights, which is required to use Call Coaching. To proceed, an admin can go to Setup →"
echo " Salesforce Go → Einstein Conversation Insights, or contact their sales representative,"
echo " to understand how to purchase ECI."
exit 1
fi
# If unknown (query rejected / no row), skip this pre-check and let the SOAP
# state detection below classify the org. TYPE_UNAVAILABLE on the SOAP read
# is the authoritative "you need Agentforce for Sales" signal.
# enableECICallScoring exposes at v68+ only.
SOAP_ENDPOINT="${INSTANCE_URL}/services/Soap/m/68.0"
# Bounds every curl call below so a hung network call can't stall the run
# indefinitely. 30s matches the Pipeline reference skill's setup-all.sh
# CURL_TIMEOUT — the only existing timeout precedent in this repo's
# SOAP-calling scripts.
readonly CURL_TIMEOUT=30
# ---------------------------------------------------------------------------
# SOAP helpers
# ---------------------------------------------------------------------------
# shellcheck source=shared/soap.sh
source "$SCRIPT_DIR/shared/soap.sh"
soap_read() {
local settings_type="$1"
local full_name="$2"
soap_call "readMetadata" "<?xml version='1.0' encoding='utf-8'?>
<soapenv:Envelope xmlns:soapenv='http://schemas.xmlsoap.org/soap/envelope/' xmlns:met='http://soap.sforce.com/2006/04/metadata'>
<soapenv:Header><met:SessionHeader><met:sessionId>${ACCESS_TOKEN}</met:sessionId></met:SessionHeader></soapenv:Header>
<soapenv:Body><met:readMetadata><met:type>${settings_type}</met:type><met:fullNames>${full_name}</met:fullNames></met:readMetadata></soapenv:Body>
</soapenv:Envelope>"
}
# classify_field / normalize_class / describe_failure now live in
# shared/soap.sh (sourced above) so install-ootb-competencies.sh's SOAP
# fallback can reuse them without duplicating this logic.
# ---------------------------------------------------------------------------
# Availability detection: read all three settings + count competencies.
# ---------------------------------------------------------------------------
echo "Detecting org state..."
echo ""
CI_RAW=$(soap_read "ConversationalIntelligenceSettings" "ConversationalIntelligence")
ECI_CLASS=$(classify_field "$CI_RAW" "enableCallCoaching")
CS_CLASS=$(classify_field "$CI_RAW" "enableECICallScoring")
ECI_ENABLED=$(normalize_class "$ECI_CLASS")
CALL_SCORING_ENABLED=$(normalize_class "$CS_CLASS")
GENAI_RAW=$(soap_read "EinsteinGptSettings" "EinsteinGpt")
GENAI_CLASS=$(classify_field "$GENAI_RAW" "enableEinsteinGptPlatform")
GENAI_ENABLED=$(normalize_class "$GENAI_CLASS")
# River Rush (Momentum) detection deferred until that arm goes live.
# Competency count via Tooling API. (sf may prepend a warning to stdout; strip.)
COMPETENCY_COUNT=""
if COMP_RAW=$(sf data query --target-org "$ORG_ALIAS" --use-tooling-api \
--query "SELECT COUNT() FROM EnablementCompetencyDef" --json 2>/dev/null); then
COMP_JSON=$(echo "$COMP_RAW" | awk '/^{/,EOF')
COMPETENCY_COUNT=$(echo "$COMP_JSON" | jq -r '.result.totalSize // empty' 2>/dev/null || echo "")
fi
echo " Einstein Conversation Insights: ${ECI_ENABLED:-unknown}"
echo " Einstein Generative AI Platform: ${GENAI_ENABLED:-unknown}"
echo " Call Coaching: ${CALL_SCORING_ENABLED:-unknown}"
echo " Configured competencies: ${COMPETENCY_COUNT:-unknown}"
echo ""
# ---------------------------------------------------------------------------
# State machine (PRD § "Behavior by org state")
# ---------------------------------------------------------------------------
# State 1a: NOT AVAILABLE — CI settings unreadable. Distinguish *why*: a
# confirmed licensing gap (TYPE_UNAVAILABLE) is a different admin action than
# a transient auth/network failure, which should be retried, not reported as
# "you need Agentforce for Sales."
if [[ -z "$ECI_ENABLED" ]]; then
case "$ECI_CLASS" in
TYPE_UNAVAILABLE:*)
echo "AI call coaching is not available in this org and they need Agentforce for Sales."
echo ""
echo " (This skill covers Einstein Conversation Insights (ECI) orgs only —"
echo " the org's ECI license appears to be missing.)"
;;
AUTH_ERROR:*)
echo "Could not check org state — authentication/permission error:"
echo " ${ECI_CLASS#AUTH_ERROR:}"
echo ""
echo " Verify the CLI session is valid (sf org login web --alias $ORG_ALIAS) and re-run."
;;
*)
echo "Could not reach Salesforce to check org state (network/timeout error)."
echo " Try again in a moment; if it persists, check the org's availability."
;;
esac
exit 1
fi
# State 4: ENABLED, competencies exist.
if [[ "$CALL_SCORING_ENABLED" == "true" && -n "$COMPETENCY_COUNT" && "$COMPETENCY_COUNT" -gt 0 ]]; then
echo "Call Coaching is already enabled — $COMPETENCY_COUNT competency record(s) exist."
echo " Ready for edit / deactivate / create-new / improvement-suggestion flows."
echo ""
echo "Next: $SCRIPT_DIR/list-competencies.sh \"$ORG_ALIAS\""
exit 0
fi
# State 3.5: ENABLED, competency count unreadable. Distinguish "we couldn't
# read the count" from "the count is genuinely zero" — falling into State 3
# tells the admin "no competencies configured" when the org may actually
# have several. Route to list-competencies to let the admin see the truth.
if [[ "$CALL_SCORING_ENABLED" == "true" && -z "$COMPETENCY_COUNT" ]]; then
echo "Call Coaching is already enabled — competency count could not be read."
echo " Ready for edit / deactivate / create-new flows."
echo ""
echo "Next: $SCRIPT_DIR/list-competencies.sh \"$ORG_ALIAS\" (to inspect the configured set)"
exit 0
fi
# State 3: ENABLED, no competencies.
if [[ "$CALL_SCORING_ENABLED" == "true" ]]; then
echo "Call Coaching is already enabled — no competencies configured yet."
echo " Ready for OOTB competency install / custom competency creation."
echo ""
echo "Next: $SCRIPT_DIR/install-ootb-competencies.sh --list (show the OOTB set, then ask OOTB vs custom)"
exit 0
fi
# State 2: AVAILABLE, NOT ENABLED — enable path.
# 2a-pre. License gate — catch unlicensed orgs BEFORE the "ECI not enabled"
# refusal below. ConversationalIntelligenceSettings returns HTTP 200 with
# enableCallCoaching=false on BOTH licensed-ECI-off and genuinely unlicensed
# orgs (the two are indistinguishable at that read), so State 1a above
# doesn't catch the unlicensed case. EinsteinGptSettings faulting with
# TYPE_UNAVAILABLE is the strongest license-gap signal we have without
# doing a write — Einstein GenAI Platform is a hard prerequisite for AI
# Call Scoring, so an org that can't even see that settings type won't be
# able to enable Call Coaching regardless of what the ECI toggle says.
#
# Not perfect: an org could have the GenAI platform present but still lack
# the specific Call Scoring license. State 2c's ECICallScoringAI
# OrgPermission check catches that residual case after GenAI is enabled.
if [[ "$GENAI_CLASS" == TYPE_UNAVAILABLE:* ]]; then
echo "AI call coaching is not available in this org."
echo ""
echo " You need Agentforce for Sales to enable AI Call Scoring."
echo " Without Agentforce for Sales, you can still set up manual call scoring"
echo " by navigating to Salesforce Go setup in your org."
exit 1
fi
echo "State: AVAILABLE, NOT ENABLED — proceeding with enable flow."
echo ""
# 2a. License gate — ECICallScoringAI OrgPermission is the authoritative
# license signal and must be checked BEFORE the "ECI not enabled" refusal
# below. ConversationalIntelligenceSettings reads enableCallCoaching=false on
# BOTH a licensed-but-toggled-off org and a genuinely unlicensed one — those
# are indistinguishable at that read. Checking here, ahead of 2b, prevents a
# no-license org from being misdiagnosed as "just flip ECI in Setup" (a
# regression this skill's eval guards against — see the NoLicense dataset).
ECICALLSCORINGAI_ENABLED=""
if ECICALLSCORINGAI_RAW=$(sf data query --target-org "$ORG_ALIAS" --use-tooling-api \
--query "SELECT SettingValue FROM OrganizationSettingsDetail WHERE SettingName='ECICallScoringAI' AND DataType='OrgPermissionSettings'" \
--json 2>/dev/null); then
ECICALLSCORINGAI_JSON=$(echo "$ECICALLSCORINGAI_RAW" | awk '/^{/,EOF')
if echo "$ECICALLSCORINGAI_JSON" | jq -e '.status == 0' >/dev/null 2>&1; then
# A missing row means "couldn't read it", NOT "license is off" — default
# to empty (unknown), never to "false". OrganizationSettingsDetail is not
# queryable (INVALID_TYPE) on some org editions (for example Developer
# Edition) and can return zero rows even when queryable; both are "can't
# tell", and collapsing them to "false" fails closed. Use select(), not
# `// empty`: SettingValue is a boolean and `//` drops an explicit false.
ECICALLSCORINGAI_ENABLED=$(echo "$ECICALLSCORINGAI_JSON" | jq -r '.result.records[0].SettingValue | select(. != null)')
fi
fi
# Only an EXPLICIT negative refuses. An unknown result — the query was
# rejected (INVALID_TYPE on this org shape) or returned no row — must NOT
# hard-fail: doing so blocks legitimately-licensed orgs where
# OrganizationSettingsDetail simply isn't queryable. This mirrors the Phase 0
# ConversationPilot gate, which defers to SOAP on any non-conclusive read. The
# 2a-pre GenAI TYPE_UNAVAILABLE gate above catches most unlicensed orgs, and
# the updateMetadata write below is the final check — a truly unlicensed org
# faults on the write. Closes the residual gap the earlier
# GenAI-proxy fix left behind.
if [[ "$ECICALLSCORINGAI_ENABLED" == "false" ]]; then
echo "AI call coaching is not available in this org and org requires Agentforce for Sales license."
exit 1
fi
if [[ -z "$ECICALLSCORINGAI_ENABLED" ]]; then
echo "Couldn't confirm the AI Call Scoring license on this org — continuing;"
echo " if the license is missing, the enable step will report it."
fi
# 2b. ECI must be enabled — this skill is ECI-only.
if [[ "$ECI_ENABLED" != "true" ]]; then
echo "Error: ECI (Einstein Conversation Insights) is not enabled."
echo " Setup → Salesforce Go → Einstein Conversation Insights → Turn ON, then re-run."
exit 1
fi
# 2c. Enable Einstein GenAI if off (PRD explicitly authorizes this).
if [[ "$GENAI_ENABLED" != "true" ]]; then
echo "Enabling Einstein Generative AI Platform (required for AI scoring)..."
GENAI_ENABLE_REQ=$(cat <<'EOF'
<?xml version="1.0" encoding="utf-8"?>
<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:met="http://soap.sforce.com/2006/04/metadata">
<soapenv:Header>
<met:SessionHeader><met:sessionId>{{ACCESS_TOKEN}}</met:sessionId></met:SessionHeader>
</soapenv:Header>
<soapenv:Body>
<met:updateMetadata>
<met:metadata xsi:type="met:EinsteinGptSettings" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<met:fullName>EinsteinGpt</met:fullName>
<met:enableEinsteinGptPlatform>true</met:enableEinsteinGptPlatform>
</met:metadata>
</met:updateMetadata>
</soapenv:Body>
</soapenv:Envelope>
EOF
)
GENAI_ENABLE_REQ="${GENAI_ENABLE_REQ//\{\{ACCESS_TOKEN\}\}/$ACCESS_TOKEN}"
GENAI_RESULT=$(classify_field "$(soap_call "update" "$GENAI_ENABLE_REQ")" "success")
if [[ "$GENAI_RESULT" == "true" ]]; then
echo " Einstein GenAI enabled"
else
case "$GENAI_RESULT" in
TYPE_UNAVAILABLE:*|false)
# Org isn't licensed for EinsteinGptSettings, or the update was
# explicitly refused with no fault and no <errors> detail — either
# way, this is a licensing gap, not a transient error, so point at
# the actual fix instead of telling the admin to just retry.
echo " Einstein Generative AI Platform cannot be enabled — $(describe_failure "$GENAI_RESULT")."
echo ""
echo " You need Agentforce for Sales to enable AI Call Scoring."
echo " Without Agentforce for Sales, you can still set up manual call scoring"
echo " by navigating to Salesforce Go setup in your org."
;;
UPDATE_REJECTED:*)
# Salesforce returned success=false with an <errors> block — surface
# the actual statusCode/message instead of guessing at a cause.
echo " Einstein GenAI enable failed: $(describe_failure "$GENAI_RESULT")"
;;
AUTH_ERROR:*)
echo " Einstein GenAI enable failed: $(describe_failure "$GENAI_RESULT")"
echo ""
echo " Verify the CLI session is valid (sf org login web --alias $ORG_ALIAS) and re-run."
;;
*)
echo " Einstein GenAI enable failed: $(describe_failure "$GENAI_RESULT")"
echo ""
echo " Try again in a moment; if it persists, check the org's availability."
;;
esac
exit 1
fi
echo ""
else
echo "Einstein Generative AI Platform is already enabled — skipping."
echo ""
fi
# Enable Call Coaching (enableECICallScoring) — skip the update if it's
# already on (state detection above only short-circuits when competencies
# also exist; an org can have enableECICallScoring=true with zero
# competencies and still reach here if that read raced with an out-of-band
# change).
if [[ "$CALL_SCORING_ENABLED" != "true" ]]; then
echo "Enabling Call Coaching..."
CS_ENABLE_REQ=$(cat <<'EOF'
<?xml version="1.0" encoding="utf-8"?>
<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:met="http://soap.sforce.com/2006/04/metadata">
<soapenv:Header>
<met:SessionHeader><met:sessionId>{{ACCESS_TOKEN}}</met:sessionId></met:SessionHeader>
</soapenv:Header>
<soapenv:Body>
<met:updateMetadata>
<met:metadata xsi:type="met:ConversationalIntelligenceSettings" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<met:fullName>ConversationalIntelligence</met:fullName>
<met:enableECICallScoring>true</met:enableECICallScoring>
</met:metadata>
</met:updateMetadata>
</soapenv:Body>
</soapenv:Envelope>
EOF
)
CS_ENABLE_REQ="${CS_ENABLE_REQ//\{\{ACCESS_TOKEN\}\}/$ACCESS_TOKEN}"
CS_RESULT=$(classify_field "$(soap_call "update" "$CS_ENABLE_REQ")" "success")
if [[ "$CS_RESULT" == "true" ]]; then
echo " Call Coaching enabled"
else
# ECI enablement is asynchronous — for a short window after the ECI toggle
# flips on, the dependent enableECICallScoring field is server-side
# read-only and updateMetadata returns INSUFFICIENT_ACCESS_OR_READONLY.
# This is a transient state that self-resolves in a few minutes, not a
# permanent failure — surface a wait-and-retry message instead of the
# generic error.
case "$CS_RESULT" in
WRITE_ERROR:INSUFFICIENT_ACCESS_OR_READONLY*)
echo " Einstein Conversation Insights is still finishing turning on."
echo " Wait a few minutes, then re-run this to enable Call Coaching."
;;
*)
echo " Call Coaching enable failed: $(describe_failure "$CS_RESULT")"
;;
esac
exit 1
fi
else
echo "Call Coaching is already enabled — skipping."
fi
echo ""
# Enable the CallCoachingReady OrgPermission — not a Metadata API type, so
# it's flipped via Tooling API PATCH on OrganizationSettingsDetail (the same
# object the OrgPermission reads above use), looked up by Id first.
CCR_LOOKUP=$(sf api request rest \
"services/data/v68.0/tooling/query/?q=SELECT+Id,SettingValue+FROM+OrganizationSettingsDetail+WHERE+SettingName='CallCoachingReady'+AND+DataType='OrgPermissionSettings'" \
--target-org "$ORG_ALIAS" 2>/dev/null) || true
# On some org editions (for example Developer Edition), OrganizationSettingsDetail
# isn't exposed at all (INVALID_TYPE) — the same unqueryable shape the Phase 3
# license gate tolerates. The authoritative enable (enableECICallScoring, flipped
# above and re-verified below) has already succeeded, so a CallCoachingReady
# lookup we physically cannot run must not fail the enable closed.
if echo "$CCR_LOOKUP" | grep -q '"errorCode"[[:space:]]*:[[:space:]]*"INVALID_TYPE"'; then
echo "CallCoachingReady OrgPermission isn't queryable on this org — skipping (Call Coaching is already enabled)."
else
CCR_COUNT=$(echo "$CCR_LOOKUP" | jq -r '.totalSize // 0' 2>/dev/null || echo 0)
if [[ "$CCR_COUNT" == "0" ]]; then
echo "Could not locate the CallCoachingReady OrgPermission record on this org."
exit 1
fi
CCR_ID=$(echo "$CCR_LOOKUP" | jq -r '.records[0].Id')
CCR_CURRENT=$(echo "$CCR_LOOKUP" | jq -r '.records[0].SettingValue')
if [[ "$CCR_CURRENT" == "true" ]]; then
echo "CallCoachingReady is already enabled — skipping."
else
echo "Enabling CallCoachingReady..."
if CCR_PATCH_RESP=$(sf api request rest \
"services/data/v68.0/tooling/sobjects/OrganizationSettingsDetail/${CCR_ID}" \
--target-org "$ORG_ALIAS" \
--method PATCH \
--header 'Content-Type: application/json' \
--body '{"SettingValue": true}' \
2>/dev/null); then
echo " CallCoachingReady enabled"
else
echo " Failed to enable CallCoachingReady"
echo " $CCR_PATCH_RESP"
exit 1
fi
fi
fi
echo ""
# 2d. Verify.
echo "Verifying..."
sleep 2
VERIFY_CLASS=$(classify_field "$(soap_read "ConversationalIntelligenceSettings" "ConversationalIntelligence")" "enableECICallScoring")
if [[ "$VERIFY_CLASS" == "true" ]]; then
echo " Confirmed: Call Coaching is now enabled"
echo ""
echo "Call Coaching enabled. Ready for OOTB competency install / custom creation."
echo ""
echo "Next: $SCRIPT_DIR/install-ootb-competencies.sh --list (show the OOTB set, then ask OOTB vs custom)"
exit 0
else
echo " Error: Verification failed: $(describe_failure "$VERIFY_CLASS")"
exit 1
fi
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |