observe new token activity
develop a structured process for monitoring newly launched solana tokens and recording relevant on-chain events.
independent on-chain observation
npcsignals
sentinel
observe first.
verify before concluding.
independent on-chain observation. structured evidence. research before conclusions.
sentinel is being developed to record early token activity, preserve evidence and support independent on-chain investigations.
npcsignals sentinel is an experimental monitoring and research system being developed to study newly launched tokens on solana.
its purpose is to collect structured observations, preserve chronological event records and help identify patterns that deserve closer investigation.
research boundary
not every unusual transaction is suspicious. not every connected wallet belongs to the same operator. not every failed token is a scam.
the objective is not automated accusations. the objective is to make better investigation possible.
controlled validation
no public feed
- version
- v0.1
- phase
- development and validation
- public access
- not available
- current priority
- reliability and evidence integrity
01
why we are building sentinel
most investigations begin after the damage is done.
a token launches. trading accelerates. wallets accumulate, distribute or exit. liquidity changes. promotional narratives evolve.
by the time questions emerge, important early activity may be buried beneath thousands of transactions.
the underlying blockchain records may remain accessible, but reconstructing what happened, in what order and under which circumstances requires time, infrastructure and careful analysis.
sentinel is our attempt to improve that process.
instead of beginning every investigation after a suspected incident, we want to develop an independent observation system capable of collecting relevant information as events unfold.
early observation is not early proof.
but a reliable chronological record can make later verification significantly more useful.
02 / what sentinel is being built to do
development objectives.
structured for investigation.
preserve chronological evidence
maintain timestamped observations that help researchers reconstruct transaction sequences and understand how activity developed over time.
support pattern investigation
build a foundation for examining token flows, wallet interactions and recurring transaction patterns without automatically assigning intent or common ownership.
distinguish evidence from interpretation
separate directly observed activity from analytical hypotheses, unresolved questions and unsupported conclusions.
make findings reproducible
work toward documented methods that allow technical findings to be independently reviewed, verified and challenged.
03
what makes sentinel different
preserve the sequence.
preserve the uncertainty.
many existing tools provide snapshots of token activity, market movements and wallet behavior.
sentinel is being developed around a different research priority: preserving structured, timestamped observations and maintaining a clear distinction between evidence and interpretation.
the system is designed to support chronological reconstruction rather than relying exclusively on what remains visible when an investigation begins.
our goal is not simply to detect unusual activity. it is to preserve what can be demonstrated, identify what remains uncertain and make subsequent investigation more effective.04 / from observation to investigation
status belongs to evidence.
not excitement.
previous npcsignals investigations, including zyrox and glitch, demonstrated the importance of reconstructing transaction sequences, tracing wallet relationships and separating verified activity from unresolved questions.
these investigations were conducted independently and completed before sentinel. they are not examples of completed sentinel operations or demonstrations of sentinel's capabilities.
observations and analytical findings are assessed using the evidence classifications described in section 07. these distinguish verified facts from supported interpretations, unresolved questions and matters outside the investigation’s scope.
a technically confirmed blockchain transaction establishes that the recorded event occurred on-chain. it does not, by itself, confirm an analytical conclusion about ownership, intent or coordination.
example / synthetic observation
illustrative data / not current system output
one event. documented without attribution.
- timestamp (utc)
- observed event
- token transfer
- source
wallet-a- destination
wallet-b- amount
- 125,000 example tokens
- transaction reference
- synthetic / no real signature
- observation status
- recorded
- evidence assessment
- illustrative / not independently verified
- operator relationship
- unresolved
05
what sentinel is not
no shortcuts to certainty.
- sentinel is not a trading bot.
- it is not a token recommendation engine.
- it does not guarantee the detection of rug pulls, market manipulation, insider activity or fraud.
- it does not establish real-world identities from blockchain connections alone.
- it does not predict token prices or guarantee protection against financial losses.
a blockchain record can demonstrate that a transaction occurred.
it cannot, by itself, establish why it occurred.
06 / development roadmap
reliability before
prediction.
-
01
current phase / development and validation
v0.1 / research prototype
establish reliable event collection, structured records, recovery procedures and technical observability.
test system stability, data completeness, operational limitations and record integrity before broader access is considered.
the immediate priority is reliability, not prediction. -
02
proposed / not committed
future research / pattern analysis
explore additional analysis of wallet interactions, transaction sequences and recurring on-chain behavior.
evaluate potential signals against historical and forward-collected evidence before making predictive claims.
any future detection or classification capabilities must demonstrate measurable performance and clearly documented limitations.
-
03
concept / subject to validation
longer-term vision / research platform
investigate whether sentinel can evolve into a publicly accessible research platform featuring searchable observations, evidence timelines, visual analysis and transparent methodology.
progress depends on technical feasibility, available resources and demonstrated research value.
interface concepts / future development
possible research interfaces.
future development.
these interface concepts illustrate possible directions for sentinel’s future development.
all displayed data is synthetic, and the interfaces are not operational.
concept a
concept preview / not a live interface
observation dashboard
a possible view of timestamped observations, activity sequences, collection status and structured event records using synthetic data.
concept b
concept preview / not a live interface
wallet relationship map
a possible analytical view of transaction paths and funding relationships. solid links represent confirmed transactions; dashed links represent possible analytical relationships. connections do not establish common ownership.
concept c
concept preview / not a live interface
evidence timeline
a possible chronological view of timestamped events, synthetic transaction references and evidence classifications including confirmed, supported and unresolved.
07 / methodology
evidence has limits.
we document them.
npcsignals follows a research-first approach. these classifications describe the status of research findings, not judgments about individuals, projects or token legitimacy.
sentinel is being developed around these same principles.- confirmed
- directly supported by reproducible evidence.
- supported
- consistent with available evidence, but not conclusively established.
- unresolved
- insufficient evidence for a defensible conclusion.
- not assessed
- outside the scope of the current investigation.
connections are not identities.
visibility is not verification.
correlation is not proof of coordination.
08 / independence and disclosure
independence requires transparency.
npcsignals separates research findings from promotional interests.
material financial interests, sponsorships and other relevant conflicts will be disclosed when applicable to published research.
research conclusions remain grounded in evidence, regardless of the project or individuals involved.
09 / corrections and attribution
addresses are not identities.
npcsignals does not attribute a wallet to a real-world individual solely because of transaction patterns, shared funding or observed connections.
material errors in published findings should be corrected transparently, with the correction and its date clearly identified.
new evidence may change an earlier assessment.
uncertainty and revisions are part of the research process, not something to conceal.10 / current development status
under development.
not publicly available.
controlled validation
sentinel is undergoing engineering work and controlled validation.
earlier internal testing has not established reliable predictive performance.
this is why the current development phase prioritizes reliable observation, data integrity and reproducible records before any predictive capability is considered.
development updates and future testing arrangements will be published on this page.
- public access
- not available
- release date
- not confirmed
- current priority
- technical reliability and evidence integrity
11
development log
internal development and validation
sentinel v0.1 remains under development.
engineering work is focused on operational reliability, controlled observation, structured event records and infrastructure validation.
public testing is not yet available.
further updates will be published when meaningful technical milestones have been independently verified.
follow development
this page is the official development record for npcsignals sentinel.
verified milestones, known limitations and future testing announcements will be published here.
sometimes the correct finding is simply: insufficient evidence.research first. conclusions later. npcsignals.com ↗