# Cost And Progress Reporting

> Produce evidence-backed run and attempt reports with evaluations, costs, risks, and promotion recommendations. Use at the end of a ClimbHill run or attempt attempt.

- Skill: `closedloop-technologies/cost-and-progress-reporting` (Agent Skill)
- Install (CLI): `npx skillmds@latest add closedloop-technologies/cost-and-progress-reporting`
- Raw SKILL.md: https://api.skillmd.com/api/skills/closedloop-technologies/cost-and-progress-reporting/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: closedloop-technologies (https://skillmd.com/u/closedloop-technologies)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/closedloop-technologies/cost-and-progress-reporting

---


# Cost and Progress Reporting

Use this skill at the end of a run or attempt attempt.

## Inputs

- Run goal
- Attempt list
- Tool logs
- Evaluation results
- Token and cost estimates when available
- Human decisions

## Procedure

1. Summarize the goal and repo state used.
2. List attempts with branch, commit, patch path, status, and summary.
3. Record tests, lint, typecheck, build, benchmark, rubric, and human review results.
4. Record costs: input tokens, output tokens, tool calls, wall-clock time, and estimated USD when available.
5. Highlight policy violations, approval-required edits, and unresolved risks.
6. Recommend promote, reject, or continue with another loop.
7. Propose next actions or GitHub issue drafts.

## Outputs

- Markdown run report
- Attempt comparison table
- Cost summary
- Risk summary
- Recommendation

## Rubric

- A maintainer can understand what happened without reading chat logs.
- Claims are grounded in recorded evidence.
- Costs and unknown costs are explicit.
- The recommendation is reviewable and reversible.