tunnel-state — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited tunnel-state (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Flagship L2 prosthetic skill. Synthesizes across the latest N summaries for a domain into a unified "save state." Every claim MUST carry [conv_id · YYYY-MM-DD] or [summary_id · date] citation — this is not decorative, it is the structural contract that distinguishes L2 from L1.
Purpose rooted in Masicampo & Baumeister: having a plan for an unfulfilled goal eliminates intrusive cognitive nagging, even without executing it. The output is the plan.
what-do-i-think)domain: str
limit: int = 10 # how many latest summaries to synthesize acrossMarkdown with structured sections, every bullet carrying a citation:
## 🧠 Tunnel State: {domain}
**Stage**: executing | refining | crystallizing | exploring
**Emotional tone**: ... [last_summary_id · date]
**Conversations summarized**: N · **Breakthroughs**: M
### Open questions
- {question} [conv_id · date]
...
### Recent decisions
- {decision} [conv_id · date]
...
### Active concepts
{concept}, {concept}, ... (flat list — concepts do not cite individually)
### Key insights
- {insight} [conv_id · date]
### Connected domains
{domain}, {domain}, ... [no individual citation]switching-cost / alignment-check)open-threads)_SUMMARIES_HINT footerIf L2 summaries are missing, degrade to keyword-expand search on L0 via _DOMAIN_KEYWORDS. Output explicitly says "operating from L0 only, citations are conversation-level" and appends _SUMMARIES_HINT.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.