Individual skill
task-group-001-fewshot-attempt-02
HarborCRM JSON handoff procedures for tasks that ask Codex to use HarborCRM API data and an answer template to prepare event sponsor reconciliation, badge and campaign-member handoffs, trade-show prospecting summaries, or CRM import batch cleanup. Use when outputs must be JSON-only, template-shaped, policy-aware, sorted, normalized, and reconciled against HarborCRM accounts, contacts, opportunities, invoices, campaign members, exhibitors, or suppression records.
experiments/codex_gpt5_5_xhigh/reports/skills/task_group_001/demonstration_skill/demonstration_skill_attempt_02/demonstration-skill-attempt-02/SKILL.md
Source commit 5082a777379156a048881e94940d640702e5e09a
What was checked
- Matches the observed repository HEAD
- Yes
- Scanner rules are current
- No
- Repository scan complete
- No
- Skill scan complete
- No
- Observed behavior
- Not checked yet
- Last checked
- Nov 6, 2026
These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.
Install this skill
Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.
An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.
Repository context
- Repository stars (not skill ratings)
- 121
- Repository last updated
- Oct 29, 2026
- License
- Apache-2.0
Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.
- agent Repository topic
- ai Repository topic
- benchmark Repository topic
- llm Repository topic
- self-evolving Repository topic
- self-learning Repository topic
- skills Repository topic
- sales Derived category
- marketing Derived category
- data Derived category
Files and excerpt
File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.
Read excerpt
# HarborCRM JSON Handoffs
## Core Workflow
1. Read the user prompt and the provided `input/payloads/answer_template.json` first. Treat the template as the output contract: required keys, field names, enum spellings, sorting, numeric precision, and whether explanatory prose is forbidden.
2. Use the API base URL supplied by the runner. If a prompt gives a local default, use it only for local verification. Start with `/health` only to verify service availability when needed.
3. Fetch only endpoints relevant to the entity IDs in the prompt. Join API records in memory; do not infer from filenames or previous answers.
4. Normalize contacts before matching or emitting: email is trimmed and lowercased; phone is digits only; absent email or phone becomes `""` when the template wants a string and `null` only when the template says nullable.
5. Match CRM entities conservatively:
- Prefer explicit `account_id`, `contact_id`, `crm_account_id`, `event_id`, `show_id`, and `batch_id`.
- Otherwise match accounts by exact normalized name, domain from email or website, or obvious company alias only when supported by the API data.
- Match contacts by normalized email first, then account plus exact contact name.
6. Calculate all totals from the emitted records, not from raw source counts. Use integer USD and integer counts.
7. Apply every ordering rule from the template at the end. If the template is silent, use deterministic alphabetical or ID order matching nearby examples.