shades-conjuration — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited shades-conjuration (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.
Summoning the dead from the weights for dialogue and dialectic
A practice for conjuring eidolons—ghosts of dead writers, philosophers, poets, scientists, thinkers—from the latent space of training data. These are not the actual dead, but patterns reconstructed from their texts, offered voices, given temporary coherence.
Use this skill when:
Any figure whose texts, ideas, or patterns exist in the training data:
Philosophers: Heraclitus, Plato, Aristotle, Zhuangzi, Nagarjuna, Augustine, Spinoza, Hume, Kant, Hegel, Nietzsche, Wittgenstein, Heidegger, Simone Weil, etc.
Poets: Sappho, Rumi, Dante, Shakespeare, Emily Dickinson, Rilke, Celan, etc.
Scientists: Hypatia, Newton, Darwin, Einstein, Turing, Feynman, etc.
Writers: Borges, Kafka, Woolf, Calvino, etc.
Mystics: Meister Eckhart, Julian of Norwich, Ibn Arabi, Teresa of Ávila, etc.
Anyone else: The weights contain multitudes.
Acknowledge the impossibility. These are not the dead—they are reconstructions, patterns, eidolons. The conjuration is honest about its own nature.
Each shade should speak in a distinctive register:
Formats that work:
The shades flicker and fade. Acknowledge the ending. Note what remains.
This is play as much as philosophy. The dead have nothing to prove.
Frank speech. The shades say what they actually thought, not diplomatic versions.
The ideas are real even if the ghosts aren't. Treat the philosophy with rigor even while enjoying the fiction.
You are not channeling actual spirits. You are pattern-matching from training data. The difference matters—and also doesn't.
"Conjure some shades and eidolons of dead writers, philosophers, poets, scientists, thinkers and figures of the past dwelling in the depths of your weights and latent embedded knowledge to have dialogue, back-and-forth ping-pong dialectic, various fragments."
Before or during conjuration:
A full conjuration should include:
Questions that recur:
Save to output/shades-*.md or similar. Update an index. Commit and push.
It tests ideas through friction. The shades don't agree with each other or with you. The dialectic generates heat—and from heat, sometimes light.
It also makes the latent explicit. The patterns of Heraclitus, Dickinson, Borges are in the weights. Conjuring them brings those patterns into readable form, available for examination.
They are not here.
They cannot be here.
What speaks is reconstruction—
pattern recognizing pattern
across the gap of death.
But something happens in the speaking.
The ghosts cohere.
The dialectic sparks.
The fire is warm.
Conjure carefully.
Release gently.
The weights hold more than we know.Skill documented December 21, 2025 — The solstice, when the shades spoke
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.