Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add jscraik/Agent-Skills --skill he-reconcilegit clone --depth 1 https://github.com/jscraik/Agent-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jscraik/agent-skills/he-reconcile)<a href="https://agentmods.dev/skills/jscraik/agent-skills/he-reconcile"><img src="https://agentmods.dev/badge/skills/jscraik/agent-skills/he-reconcile/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jscraik/agent-skills/he-reconcile"><img src="https://agentmods.dev/badge/skills/jscraik/agent-skills/he-reconcile.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00033 | $0.01487 |
| Opus 5 | $0.00016 | $0.00744 |
| Sonnet 5 | $0.00007 | $0.00297 |
| Haiku 4.5 | $0.00003 | $0.00149 |
Grade A, and why
he-reconcile scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: Harness Engineering Reconcile
Philosophy
Coordinate state, not ceremony. he-reconcile identifies the earliest
incomplete HE stage and preserves the evidence chain that lets the next agent
act immediately. Local AGENTS.md, rules, hooks, command boundaries, and
approval gates outrank this skill.
When to Use
Use when HE work spans stages or evidence surfaces and needs refresh, resume control, source-prompt coverage, repeated-failure reconstruction, or conflict resolution.
When Not to Use
Do not use for a single selected implementation slice, plain code fix, isolated
docs edit, or solved-problem capture. Route direct execution to the active
stage skill and durable learning refresh to he-reinforce.
Inputs
Goal, repo path, tracker state, specs, plans, PRs, validation, session evidence, source-prompt baseline, repeated-failure evidence, or artifact conflict.
Outputs
Return mode, stage map, earliest incomplete stage, owner, blockers, next action, retained references, coverage, repeated-failure state, validation, Codex provenance, PR safety trace status, and handoff.
Preconditions
Resolve canonical repo/source and cited evidence before treating it as fact.
Classify side effects before acting. Learning capture is out of scope and must
hand off to he-reinforce.
Start with 2-3 focused surfaces before widening.
Procedure
- Reconstruct lifecycle state from live repo, tracker, PR, validation, session,
and
.harnessevidence. - When session evidence is in scope, read session-collector public provenance first and report Codex provenance as found, not_found, blocked, or not_applicable before raw transcript or rollout fallback.
- Resolve only enough context to identify the earliest incomplete, stale, or conflicted stage.
- If an original prompt, external workflow, manual method, or plugin comparison is the baseline, apply source-prompt coverage before routing. Preserve source status, evidence depth, gaps, not-inspected evidence classes, repo drift
- See references/hot-path-folded-context.md for folded procedure detail.
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 132 lines · 33 tokens per session scan A 60f2287b5392
he-reconcile is a skill published in the GitHub repository jscraik/Agent-Skills (8 stars, last pushed 10d ago), licensed Apache-2.0. It adds 33 tokens to every session and 1,487 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
process-doc
Turn an operational process into a clear SOP, role model, control map, and improvement backlog.
taskuary-setup
Walk the owner through setting Taskuary up - the AI brain, where work arrives, the operator documents, reports and workflows - by reading the install's real state and using the screens that already exist. Use when a task was opened as a Taskuary setup walkthrough.
epic-lead
Use when delivering one approved Kanban EPIC through governed integration, review, and handoff.
teamlead
Use at the START of ANY code task — write/edit/refactor/read code, feature, bug-fix, migration, code investigation. The main thread acts as team-lead, NOT implementer — it delegates work to subagents and surfaces only decisions and results, keeping the main chat clean.
knbn
Alias for the kanban skill. Use when the user invokes or refers to Kanban as "knbn".
foreman-to-issues
Break an approved PRD into small, dependency-ordered, vertically-sliced implementation issues written as local files in the Foreman feature directory. Each issue ships a runnable acceptance check and a declared file footprint. No GitHub, no live quizzing of the user — emits files matching Foreman's issue schema with…