Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add panayiotism/claude-harness/plugin install claude-harnessWrote 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/panayiotism/claude-harness/start)<a href="https://agentmods.dev/skills/panayiotism/claude-harness/start"><img src="https://agentmods.dev/badge/skills/panayiotism/claude-harness/start.svg" alt="Measured on agentmods" 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 | $0.00045 | $0.01946 |
| Opus 5 | $0.00023 | $0.00973 |
| Sonnet 5 | $0.00009 | $0.00389 |
| Haiku 4.5 | $0.00005 | $0.00195 |
Grade A, and why
start 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 — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Start - Session Initialization and Context Compilation
Prepare for a new coding session: compile memory context, show status, sync GitHub.
Session ID: ${CLAUDE_SESSION_ID} — session state lives in .claude-harness/sessions/${CLAUDE_SESSION_ID}/.
Current Harness Snapshot (auto-compiled at invocation)
!python3 "${CLAUDE_PLUGIN_ROOT}/scripts/compile-briefing.py" .claude-harness 2>/dev/null || echo "(harness state not initialized - run /claude-harness:setup)"
Phase 0: Migration Check
Migrations are handled by setup.sh, which the SessionStart hook runs automatically whenever the plugin version changes. Only act here if legacy files are visible at the project root (feature-list.json, feature-archive.json, claude-progress.json, working-context.json, agent-context.json, agent-memory.json, init.sh): run bash {plugin-root}/setup.sh once (plugin root is in the session context) and report what it migrated.
Phase 0.5: Set Paths
- Set path variables:
FEATURES_FILE=".claude-harness/features/active.json"ARCHIVE_FILE=".claude-harness/features/archive.json"MEMORY_DIR=".claude-harness/memory/"SESSION_DIR=".claude-harness/sessions/${CLAUDE_SESSION_ID}/"
Important: All subsequent phases must use these path variables instead of hardcoded paths.
Phase 1: Context Compilation (Memory System)
The snapshot above already contains the compiled overview (features, recent decisions, failures to avoid, learned rules, last session). Use it directly - do NOT re-read the memory files unless you need detail beyond the snapshot.
- Deep-dive reads (only when needed): if the active feature requires filtering memory by
relatedFiles, read IN PARALLEL (single message, multiple Read calls) -- memory layers are OKF concept files (markdown with YAML frontmatter), so read each directory'sindex.mdfirst, then only the relevant concept files:${FEATURES_FILE}(full feature entries)${MEMORY_DIR}/failures/index.md(then relevant Failure concepts)${MEMORY_DIR}/successes/index.md(then relevant Success concepts)${MEMORY_DIR}/decisions/index.md(then relevant Decision concepts)${MEMORY_DIR}/rules/index.md(then relevant Rule concepts)
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 · 144 lines · 45 tokens per session scan A 680baad07ac5
start is a skill published in the GitHub repository panayiotism/claude-harness (79 stars, last pushed 1mo ago), licensed MIT. It adds 45 tokens to every session and 1,946 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 2026-08-30.
Other skills, from other repositories
conventional-commit
Workflow for generating conventional commit messages following the Conventional Commits specification. MUST be invoked every time a commit is created. Guides construction of standardized commit messages with correct type, scope, description, body, and footer.
python-release
Handle Python SDK release, build, bump, packaging metadata, PyPI client pin, uv.lock, nox/build workflow, and publish verification changes. Use for Python release process work or dependency pin bumps; do not use for ordinary Python feature implementation.
release-notes
Generate and publish concise, evidence-based notes in the body of the latest existing GitHub Release. Use only when the user explicitly invokes $release-notes or explicitly asks to update the latest existing GitHub Release body. Do not invoke for general release planning, changelog, tag, or version tasks.
contributor
End-to-end open source contribution workflow: from scanning issues to submitting PRs. Use this skill whenever the user wants to contribute to an open source project, find issues to fix, submit a pull request, fork a repo to contribute, fix a GitHub issue, or mentions 'open source contribution'. Also trigger when they…
review-gate
Use before an agent-produced diff is committed, pushed, opened as a PR, merged, landed, or applied to user files when explicit implementation approval is missing. Trigger for review gate, review pack, approve before landing, diff first then land, human approval, merge gate, commit gate, push gate, or PR readiness.…
git-commit-smart
Generates meaningful, conventional commit messages automatically. Use when committing code changes. Analyzes diff to create descriptive commits following best practices.