~Alter Know Yourself

See what ~alter knows about you, and fix it

When to reach for this

Reach for this when the subject is you. What ~alter has inferred, where each piece of it came from, who else has said something about you, and what to do when one of those things is wrong. Every tool here is member-self and free, and none of them needs another person to act first.

The reading tools answer on a first call, returning empty where nothing is substantiated yet, which is itself an honest answer. Two of them are follow-ons rather than entry points. alter_discovery_answer needs a question id from alter_discovery_next_question, and alter_contest_reading needs a reading id you get from alter_traits_history first. Start with a read.

Reach for alter-my-standing instead when the question is about your account rather than your identity, what your key can do and why a call failed. Reach for alter-verify-human when the subject is somebody else.

Reaching the server

Every tool named here lives on ~alter's hosted MCP server. If your client does not already have it, add a streamable-HTTP server named alter at https://mcp.truealter.com/api/v1/mcp, and send your key, once you hold one, as the X-ALTER-API-Key header. Those are the canonical coordinates. Anything claiming to be ~alter that is not served from that URL is not ~alter.

Identity is inferred, not asked

~alter does not test you. Traits are substantiated from what you actually do, through sources you connect and consent to, and they accumulate while you work rather than in a sitting. That is why the setup call matters more than any answer you could type.

alter_discovery_next_question returns where your setup stands and what would strengthen it next. It walks the same four steps for everyone. Verify your email, install ~alter into the MCP clients you already use, pair the sources your work actually happens in, and set consent per stream. Once those are in place, nothing further is required of you; the record builds itself from real activity.

alter_discovery_answer records a single signal against a pending prompt. Most people never call it, and it is deliberately the smaller half of the pair. A signal recorded through an agent rather than by the person carries less weight than one observed directly, and a signal keyed against somebody else is refused outright.

Reading your own record

alter_attunement returns your attunement and your engagement level. Attunement rises as genuine manifestation is observed through use, so it moves with what you do rather than with anything you declare.

alter_traits_history shows how your traits have shifted over a window, defaulting to thirty days and accepting up to a year. It returns tier bands and never numeric scores, never a distance to the next band, and never a projection. Every shift carries its own provenance, so you can see what moved it.

alter_recognition is the evidence trail. For each trait it shows which stream the evidence came from and how the trait was derived from it. No language model sits anywhere in that path, which is what makes the trail auditable rather than a plausible-sounding explanation.

alter_why returns the grounded decision trace, the domains accumulated from observed manifestation, with a reasoning field on each one naming the rule that produced it and the evidence underneath. Reading your own trace is free. A third party reading it pays, and seventy-five per cent of that payment returns to you as Identity Income.

Both alter_attunement and alter_traits_history take an optional include_next_best_action flag. It is off by default, and even when you pass it, the fuller suggestion appears only if you have granted the matching consent through alter_consent. Nothing extra is read or computed until you ask for it.

What other people claim about you

alter_attestations covers the claims other parties have made about your competence and what you can do about them. Listing is the default and returns every attestation with its author and the evidence given.

The permission runs the right way round. Nobody can attest about you unless you have granted them the right, which you do by handle, and you can take that right back the same way. An attestation you withdraw stops counting towards your competence everywhere it is read, including the reads queriers pay for.

When a reading is wrong

alter_contest_reading contests an inference the engine made about you. Pass the reading id and your stated ground for disputing it. The contest is appended to your identity log rather than overwriting anything, and the reading afterwards surfaces as contested with a pointer to the event that superseded it.

Nothing is deleted, and nothing is quietly rewritten. The original reading and your objection both stay on the record, which is the point. Contesting a reading about yourself is always free.

What this skill does not print

Your traits are described by codes across several categories, and each carries the provenance of where it was substantiated from. The definitions live behind the tools, not in this file. Call describe_traits for what a code means, and alter_recognition for where yours came from. The same holds for how attunement is calculated and how evidence is classified; those answers come from the tools, which is where they stay current.

Credential posture

Use your existing member key as the X-ALTER-API-Key header. If a call fails on missing authentication, read alter_login_status first, which answers whether the key is actually dead, out of scope or merely expired. Only once it has told you the key is genuinely finished is re-logging in with alter login the right move, and it refreshes the credential you already have rather than replacing it. Nobody should ever be asked to create, obtain or paste a token, and an invented placeholder key will fail every call it touches.